Scaffold Role of a Mitogen-activated Protein Kinase Phosphatase, SKRP1, for the JNK Signaling Pathway*

Stress-activated protein kinase (SAPK) pathway-regulating phosphatase 1 (SKRP1) has been identified as a member of the mitogen-activated protein kinase (MAPK) phosphatase (MKP) family that interacts physically with the MAPK kinase (MAPKK) MKK7, a c-Jun N-terminal kinase (JNK) activator, and inactivates the MAPK JNK pathway. Although these findings indicated that SKRP1 contributes to the precise regulation of JNK signaling, it remains to be elucidated how SKRP1 is integrated into this pathway. We report that SKRP1 also plays a scaffold role for the JNK signaling, judged by the following observations. SKRP1 selectively formed the stable complexes with MKK7 but not with MKK4 and biphasically regulated the MKK7 activity and MKK7-induced gene transcription in vivo. Co-precipitation analysis between SKRP1 and MKK7-activating MAPKK kinases (MAPKKKs) revealed that SKRP1 also interacted with the MAPKKK, apoptosis signal-regulating kinase 1 (ASK1), but not with MAP kinase kinase kinase 1 (MEKK1). Consistent with these findings, SKRP1 expression increased the ASK1-MKK7 complexes in a dose-dependent manner and specifically enhanced the activation of MKK7 by ASK1. Thus, our findings are, to our knowledge, the first evidence to show that an MKP also functions as a scaffold protein for the particular MAPK signaling.

Although different MAPKs are regulated by distinct signaling modules, they can be activated by a subset of the same kinases that are used in more than one signaling pathway. Therefore, mechanisms to ensure signaling specificity are required for proper response to distinct stimuli. One emerging mechanism is the formation of multienzyme complexes through scaffold proteins (56 -61), which simultaneously associate with several components of the particular MAPK pathway. To date, distinct mammalian scaffold proteins, such as JIP1 (62,63), JSAP1 (64), MP1 (65), KSR (66), and ␤-arrestin 2 (67), have been identified for MAPK pathways and appear to contribute to the precise regulation of these pathways by forming multienzyme complexes with several kinases. However, it has not been elucidated how dual specificity MAPK phosphatases (MKPs) can be integrated into MAPK signaling complexes or interact selectively with these complexes. SKRP1 (68), a member of the MKP family, negatively regulates the activation of the JNK pathway, consistent with the finding that it is highly specific for JNK in vitro. Although SKRP1 does not bind directly to its target JNK, it specifically co-precipitates with MKK7. Furthermore, SKRP1 and JNK bind to different regions on MKK7. These findings indicate that SKRP1 interacts indirectly with its physiological target JNK through MKK7, thereby leading to the specific regulation of the JNK signaling pathway. Here we report that SKRP1 also functions as a scaffold protein for the JNK signaling, judged by the following observations as reported previously (56 -62, 64, 65, 67, 69, 70): 1) selective interaction with several components of the signaling module, 2) enhancement of signaling by the boundary formation of protein kinases, and 3) selective inhibition of signaling by the scaffold overexpression. Actually, SKRP1 specifically interacted with MKK7 and ASK1, and consistent with these findings, its expression selectively enhanced the ASK1-induced MKK7 activation. Furthermore, the inhibition of MKK7 activity associated with SKRP1 overexpression was also found.
Kinase Assays-Kinase assays were as described previously (68). Briefly, immunoprecipitations or glutathione-Sepharose precipitations from cell lysates were carried out using appropriate antibodies together with protein G-Sepharose beads (Amersham Biosciences) or using glutathione-Sepharose 4B (Amersham Biosciences), respectively. After incubation for 3 h at 4°C, the complexes were processed for in vitro kinase assays using a substrate protein (3.0 g). For coupled protein kinase assays, 2.0 g of MAPKs (His-JNK2 and His-p38␣) and 5.0 g of MAPK substrates were used.
Reporter Gene Assays-NIH3T3 cells (1.0 ϫ 10 5 cells) were transiently transfected with the reporter plasmid pFR-Luc (Stratagene) and the indicated expression plasmids together with pRL-TK (Promega) to measure the transfection efficiency. Expression plasmids were transfected at the following amounts: 200 ng of pFR-Luc, 12.5 ng of pFA-ATF2 (Stratagene), and 200 ng of pRL-TK; 0, 62.5, 125, 250, or 500 ng of Myc-SKRP1; 125 ng of HA-MKK7; 125 ng of HA-JNK2. The total amount of DNA (1162.5 ng) was kept constant by supplementation with empty vector DNAs. After culture for 48 h, luciferase assays were performed using the dual-luciferase reporter system (Promega), in which relative firefly luciferase activities were calculated by normalizing transfection efficiency according to the Renilla luciferase activities. The relative firefly luciferase activities detected in the cell lysates are presented. The data shown are the mean Ϯ S.D. (n ϭ 3).

Selective Interaction of SKRP1 with the MAPKK MKK7-
Association of SKRP1 with the MAPKK MKK7, but not MKK4, was detected in the quiescent state of cells (68). Therefore, we examined whether the activation status of the MAPKKs, MKK4 or MKK7, affects the interaction with SKRP1. Because previous reports (13)(14)(15)(16)18) showed that both MKK4 and MKK7 were activated in response to anisomycin, either HA-MKK4 or HA-MKK7 was co-expressed in COS-7 cells together with GST-SKRP1 and treated with anisomycin, followed by co-precipitation analysis using glutathione-Sepharose as described previously (13,62,67,(73)(74)(75)(76). This study showed that SKRP1 interacted with MKK7, but not with MKK4, and their selective interaction was detected in both quiescent and activated cells (Fig. 1).
Regulation of the MKK7 Signaling by SKRP1-To explore further the relevance of the interaction of SKRP1 and the MKK7-JNK pathway, we examined the effect of SKRP1 on the MKK7 activity and the activation of transcription by this pathway in mammalian cells. COS-7 cells were transfected with HA-MKK7 and GST-JNK2 together with varying amounts of Myc-SKRP1, and both kinases precipitated from the cell lysates were processed for in vitro kinase assays. Coupled protein kinase assays demonstrated that SKRP1 biphasically regulated the phosphorylation of JNK2 by MKK7 ( Fig. 2A, top) but did not cause any change in the phosphorylation of JNK2 by MKK4 (data not shown). Furthermore, in parallel to the activity of MKK7, SKRP1 biphasically regulated that of JNK2 induced by MKK7 ( Fig. 2A, 2nd panel). Next, we tested whether SKRP1 suppresses the MKK7 catalytic activity by dephosphorylating the critical residues of MKK7 required for its activation. We therefore examined whether the SKRP1 catalytically inactive mutant (SKRP1 C149S) also inhibits the MKK7 activity. HA-MKK7 was co-expressed in COS-7 cells together with SKRP1 C149S at the same amounts as when the inhibitory effect of SKRP1 was detected as shown in Fig. 2A, followed by a coupled protein kinase assay. This study showed that the inhibitory effect was also detected when SKRP1 C149S was expressed together with MKK7 ( Fig. 2B, top), suggesting that SKRP1 phosphatase activity was not required for this inhibition.
Furthermore, as shown in Fig. 1, we observed the stable interaction of SKRP1 with both nonphosphorylated and phosphorylated forms of MKK7 (Fig. 2B, 2nd panel). These findings raised the possibility that SKRP1 disturbs the physical inter- HA-MKK7 was immunoprecipitated from the cell lysates, and the kinase activity was examined by a coupled protein kinase assay using recombinant His-JNK2 and GST-cJun as substrates. The phosphorylated His-JNK2 and GST-cJun were detected after SDS-PAGE by autoradiography (top), and phosphorylated His-JNK2 was then quantitated (right panel). GST-JNK2 was also precipitated from the cell lysates using glutathione-Sepharose and processed for in vitro kinase assays using GST-cJun as a substrate (2nd panel). The amounts of Myc-SKRP1 (3rd panel), HA-MKK7, and GST-JNK2 (bottom) in the cell lysates were examined by immunoblotting (WB). Similar results were obtained in three different experiments. B, inhibition of MKK7 activity by SKRP1 is phosphatase activity-independent. COS-7 cells were transfected with the indicated plasmids and treated as described in Fig. 1. HA-MKK7 immunoprecipitated from the cell lysates was then processed for a coupled protein kinase assay, and the phosphorylated His-JNK2 and GST-cJun were detected after SDS-PAGE by autoradiography (top) and quantitated (right panel). Coprecipitation analysis was also performed as described in Fig. 1. C, effect of SKRP1 overexpression on the MKK7-induced ATF2 transcriptional activity. Either control or HA-MKK7 was transiently transfected into NIH3T3 cells together with pFR-Luc and increasing amounts of Myc-SKRP1, as indicated. Transfection efficiency was monitored by co-transfection with pRL-TK. After culture for 40 h, the cells were left untreated or treated with thapsigargin (0.5 M) and incubated for 8 h. Reporter gene assays in the cells were then performed to measure the transcriptional activity of a GAL4-ATF2 fusion protein as described under "Experimental Procedures." The data shown are the mean Ϯ S.D. (n ϭ 3). action between MKK7 and its substrate JNK and thereby inhibits the MKK7-mediated JNK2 phosphorylation. However, because SKRP1 did not compete with JNK2 for the interaction with MKK7 (68), we ruled out this possibility.
Finally, because the JNK pathway is known to regulate the transcription factor ATF2 (77-79), we investigated the effect of SKRP1 expression on the MKK7-induced ATF2 transcriptional activity in either untreated or stress-activated cells. In vitro kinase assays showed that SKRP1 strongly inhibited both TNF␣-and thapsigargin-induced JNK activation (68). We therefore performed transient reporter-gene assays in untreated or thapsigargin-treated cells, and we found that SKRP1 suppressed or enhanced MKK7-induced ATF2 activity (Fig.  2C), consistent with the result in Fig. 2A. Thus, overexpression of SKRP1 biphasically regulated signaling through the MKK7-JNK pathway.
Association of SKRP1 with the MAPKKK ASK1-SKRP1 overexpression within the limited range suppressed the phosphorylation of JNK2 by MKK7, raising the possibilities as described above. However, we ruled out these possibilities because of our observations. Here we found that SKRP1 overexpression also enhances the MKK7-induced JNK2 phosphorylation. The finding indicated that SKRP1 interacts with an upstream activating kinase of MKK7 and regulates its activity. Thus, we examined whether SKRP1 interacts with an MKK7activating kinase. Furthermore, it has been reported (56,62,66,69,70) that when one functional component of the signaling complex is overexpressed, the signaling pathway is inhibited by titrating out each individual component of the complex. We therefore examined the possibility that inhibition of MKK7-JNK signaling associated with SKRP1 overexpression reflects its function of sequestering an upstream activating kinase from MKK7. The MAPKKKs ASK1 (32), MEKK1 (23), MLK3 (28), and DLK (29) have been identified as just such activating kinases of MKK7 (13,20,80). However, two members of the MLK group, MLK3 and DLK, are also known to interact with MKK7 and JNK through scaffold proteins JIP1 (62), JIP2 (81), or JIP3 (75). Therefore, we tested whether SKRP1 can interact with the other MAPKKKs, ASK1 or MEKK1. In this experiment, GST-SKRP1 was co-expressed in cells together with either HA-ASK1 or HA-MEKK1, followed by co-precipitation analysis. This study demonstrated that ASK1 was detected in SKRP1 precipitates, although the association of SKRP1 with MEKK1 was not detected (Fig. 3A). Furthermore, because ASK1 is activated by TNF␣ (72), we examined the effect of TNF␣ on the ASK1-SKRP1 interaction, and their interaction was also found in TNF␣-stimulated cells (Fig. 3B).
Facilitation of the ASK1-MKK7 Interaction by SKRP1-Our study shows that both MKK7 and ASK1 were detected in SKRP1 precipitates. These data suggest that SKRP1 may serve as one functional component to facilitate specific protein-protein interactions within the JNK signaling complex. To test this directly, we assessed the effect of varying amounts of SKRP1 expression on the association of MKK7 with ASK1. The study showed that expression of SKRP1 enhanced the co-precipitation of MKK7 with ASK1 (Fig. 4). Thus, the result indicated that SKRP1 can facilitate the ASK1-MKK7 interaction to form the JNK signaling complex.
Enhancement of the ASK1 Activity by SKRP1 Expression-SKRP1 overexpression within the limited range inhibited the MKK7 activity ( Fig. 2A, first 3 lanes), in part, by sequestering an upstream activating kinase ASK1 from MKK7 as described above, whereas SKRP1 at higher expression enhanced the MKK7 activity ( Fig. 2A, 4th lane). Furthermore, we found the physical interaction of ASK1 and SKRP1 in vivo (Fig. 3). Thus, the association of SKRP1 with ASK1 raised the possibility that SKRP1 overexpression by itself may contribute to the up-regulation of ASK1 activity. Therefore, to examine the effect of SKRP1 expression on the ASK1 activity, HA-ASK1 was cotransfected in NIH3T3 cells together with either control or Myc-SKRP1, followed by an in vitro kinase assay. As expected, the study showed that both SKRP1 and SKRP1 C149S enhanced the basal and TNF␣-induced ASK1 activities ( Fig. 5 and data not shown). Together, our findings indicated that the  4. SKRP1 increases the ASK1-MKK7 complexes. GST (control) and GST-MKK7 were transiently transfected into COS-7 cells together with HA-ASK1 and increasing amounts of Myc-SKRP1, as indicated. Glutathione-Sepharose precipitates and cell lysates were examined by immunoblotting using the indicated antibodies. It was also verified that GST (control) was precipitated using glutathione-Sepharose beads (data not shown). Similar results were obtained in two different experiments. inhibition of MKK7 activity and enhancement of ASK1 activity, associated with SKRP1 overexpression, contribute to the regulation of total catalytic activity of MKK7. Thus, the result as shown in Fig. 2A may be explained by this idea.
SKRP1-dependent Enhanced Activation of MKK7 by ASK1-Next, to examine the effect of SKRP1 on the ASK1-MKK7 pathway, either GST-MKK7 or GST-MKK3 was co-expressed in COS-7 cells together with HA-ASK1 and increasing amounts of Myc-SKRP1, followed by coupled protein kinase assays for GST-MAPKKs. These studies revealed that SKRP1 expression enhanced the MKK7 activation by ASK1 (Fig. 6A). In contrast, SKRP1 did not enhance the MKK3 activation by ASK1 (Fig.  6B), although ASK1 is also known to activate MKK3 (32). Therefore, these findings indicated that SKRP1-dependent MKK7 activation by ASK1 does not reflect the ASK1 activity potentiated by co-expression with SKRP1 as described in Fig. 5 but a scaffold role of SKRP1 for the ASK1-MKK7 pathway. Furthermore, the above observations are consistent with our results in Fig. 4.
Effect of SKRP1 Expression on the ASK1-mediated JNK2 or p38␣ Basal Activity-Our results described above suggested that SKRP1, a member of the MKP family, can also function as a scaffold protein for the JNK signaling in vivo. We therefore examined how SKRP1 regulates the ASK1/MKK7-mediated JNK2 activity. GST-JNK2 was co-expressed in COS-7 cells together with HA-ASK1, HA-MKK7, and increasing amounts of Myc-SKRP1, and the activity of GST-JNK2 was examined. This study showed that SKRP1 expression had no stimulatory effect on the basal JNK2 activity (Fig. 7A), consistent with the result that JNK2 does not bind directly to SKRP1 (68). Furthermore, SKRP1 failed to inhibit the basal JNK2 activity mediated by ASK1 (Fig. 7A), whereas the inhibitory effect of SKRP1 on the basal p38␣ activity was detected in a SKRP1dependent manner (Fig. 7B).
Selective Inhibition of the Activation of the TNF␣-induced ASK1-mediated JNK2 Pathway by SKRP1-To date, it has been shown that ASK1 and MKK7 are involved in the TNF␣induced JNK activation (13,14,72). Therefore, to explore further the physiological significance of SKRP1, we investigated the effect of SKRP1 on the TNF␣-induced JNK activity in the ASK1-MKK7-JNK2 pathway. Next, to examine whether SKRP1 selectively recognizes the TNF␣-induced JNK activation mediated by ASK1, 293 cells were transfected with SKRP1 and MKK7 in the absence or presence of ASK1 and then treated with TNF␣. In vitro kinase assays showed that in cells without co-expression of ASK1, SKRP1 had no inhibitory effect on the TNF␣-induced JNK2 activation (Fig. 8A). However, in cells also expressing ASK1 together with SKRP1 and MKK7, SKRP1 inhibited the TNF␣-induced JNK2 activation in a dosedependent manner (Fig. 8B). These findings indicated that after treatment with TNF␣, SKRP1 recognizes and inactivates the JNK pathway mediated by ASK1 in vivo.
In the present study, we examined how SKRP1, a member of such an MKP family, is integrated into the MAPK JNK pathway. We first found that SKRP1 specifically interacted with the MAPKK MKK7, and their interaction was also detected in both quiescent and activated cells (Fig. 1). Furthermore, our study showed that SKRP1 expression within the limited range suppressed the phosphorylation of JNK2 by MKK7 ( Fig. 2A). Thus, our findings presented above raised the following possibilities. 1) SKRP1 dephosphorylates the critical residues of MKK7 required for its activation and thereby suppresses the MKK7 catalytic activity. 2) SKRP1 disturbs the physical interaction between MKK7 and its substrate JNK and inhibits the phosphorylation of JNK2 by MKK7. 3) SKRP1 competes for ATP on the ATP-binding site of MKK7 to inhibit the MKK7 kinase activity. 4) SKRP1 suppresses the MKK7 catalytic activity by inactivating the MKK7-activating MAPKKKs. 5) SKRP1 inhibits the MKK7 catalytic activity by sequestering upstream activating kinases from MKK7. Therefore, we examined the validity of each of them.
First, as shown in Fig. 2B, the phosphorylation of JNK2 by MKK7 was inhibited by SKRP1 C149S as well as SKRP1. SKRP1 C149S has been identified as a catalytically inactive mutant, because it was unable to dephosphorylate an artificial substrate p-nitrophenyl phosphate and phosphorylated JNK2 in vitro (68). Thus, our results suggested that the phosphatase activity of SKRP1 is not required for its suppression of MKK7 catalytic activity, and therefore we ruled out the first possibility.
The JNK2-interacting domain on MKK7 has been reported to be located at the N terminus (13, 86) and did not overlap with the observed interaction domain of SKRP1 on MKK7 (68). Therefore, these findings excluded the second possibility that SKRP1 competes with JNK for the same domain on MKK7 to inhibit the phosphorylation of JNK by MKK7. As shown in Fig.  2A, we also found that SKRP1 expression resulted in both inhibition and enhancement of the JNK phosphorylation by MKK7. These findings suggested that the catalytic activity of MKK7 is biphasically regulated by SKRP1 expression. Thus, it is unlikely that SKRP1 blocks the ATP binding on MKK7 and thereby inhibits its ATP-transferring activity (the third possibility). Furthermore, the fourth possibility was also excluded because of the same reason. Actually, to the direct negation of this possibility, the activity of ASK1, one of the MKK7-activating MAPKKKs, was enhanced by co-expression of SKRP1 (Fig. 5).
Finally, all our data presented above are consistent with the fifth possibility that SKRP1 overexpression suppresses the MKK7 kinase activity by sequestering an upstream activating kinase from MKK7. This idea is supported by our observations that both SKRP1 and its catalytically mutant SKRP1 C149S interact physically with MKK7 ( Fig. 1) or ASK1 (Fig. 3B) and also suppress the phosphorylation of JNK2 by MKK7 (Fig. 2B).
The dose-response analysis of SKRP1 using a coupled protein kinase assay revealed that as the amounts of SKRP1 expression altered, SKRP1 negatively or positively regulated the MKK7-induced JNK phosphorylation. Similar results were obtained from the reporter-gene assays using NIH3T3 cells (Fig. 2C), in which the effect of SKRP1 on the MKK7-induced gene transcription was examined. In contrast, such a doseresponse effect was not observed in in vitro experiments using recombinant SKRP1 and MKK7 proteins (data not shown), suggesting that SKRP1 also modulates the MKK7 signaling pathway at the points other than MKK7 or JNK in vivo. Consistently, as already described, SKRP1 expression by itself enhanced the ASK1 activity (Fig. 5). Although not investigated in detail, one explanation for the mechanism by which SKRP1 enhances the ASK1 activity is that SKRP1 facilitates the oligomerization of ASK1, which has been reported to up-regulate its activity (87). Because the higher amounts of SKRP1 can hinder the interaction between ASK1 molecules, the idea is supported by our findings that as SKRP1 expression increases, the ASK1 activity displays an initial increase and then starts to decrease (data not shown). To date, it has been proposed that when one particular component of a given signaling complex is overexpressed, it titrates out other components that interact with it and leads to the down-regulation of total throughput of signaling complex (56,62,66,69,70). Several members of the TRAF family proteins, TRAF2 (88), TRAF5 (89,90), and TRAF6 (91,92), are reported to interact with ASK1 and upregulate its activity through their overexpression (72). Therefore, to explore the functional relationship between SKRP1 and these TRAF family members, we performed co-precipitation analysis to examine their interactions, but we could not find co-precipitation in detectable levels.
As observed in the present study, the selective interaction of SKRP1 with the MAPKK MKK7 and the MAPKKK ASK1 led us to examine whether SKRP1 functions as a scaffold protein for the JNK signaling pathway. We therefore tried to confirm the identification of SKRP1 as a scaffold that 1) interacts selectively with several components of the particular signaling module (as described above), 2) enhances the particular signaling by the boundary formation of protein kinases (see below), and 3) selectively inhibits the particular signaling by its overexpression (see below). Thus, we further examined the effect of SKRP1 expression on the association of MKK7 with ASK1, and we found that SKRP1 facilitated their association (Fig. 4).
Consistent with this finding, SKRP1 also specifically enhanced the activation of MKK7 by ASK1 (Fig 6A). Furthermore, as shown in Fig. 2A, overexpression of SKRP1 suppressed the MKK7 activity, and as described previously (56,62,66,69,70), this inhibition is supposed to reflect a scaffold role of SKRP1. Thus, although the precise mechanism awaits further detailed study, all our results support a scaffold role of SKRP1 for the JNK pathway. To date, distinct mammalian scaffold proteins, such as JIP1 (62,63), JSAP1 (64), MP1 (65), KSR (66), and ␤-arrestin 2 (67), have been identified for MAPK pathways by forming multienzyme complexes with several kinases. However, to our knowledge it has not been reported so far that an MKP plays a scaffold role for the MAPK pathway. Our findings thus provide the first evidence that an MKP, which has been known as one major negative regulator of the MAPK signaling module, also acts as a scaffold protein in their corresponding MAPK pathways.