Conditional Expression of the Mitogen-activated Protein Kinase (MAPK) Phosphatase MKP-1 Preferentially Inhibits p38 MAPK and Stress-activated Protein Kinase in U937 Cells*

Phorbol ester tumor promoters, such as phorbol 12-myristate 13-acetate (PMA), are potent activators of extracellular signal-regulated kinase 2 (ERK2), stress-activated protein kinase (SAPK), and p38 mitogen-activated protein kinase (MAPK) in U937 human leukemic cells. These kinases are regulated by the reversible dual phosphorylation of conserved threonine and tyrosine residues. The dual specificity protein phosphatase MAPK phosphatase-1 (MKP-1) has been shown to dephosphorylate and inactivate ERK2, SAPK, and p38 MAPK in transient transfection studies. Here we demonstrate that PMA treatment induces MKP-1 protein expression in U937 cells, which is detectable within 30 min with maximal levels attained after 4 h. This time course coincides with the rapid inactivation of PMA-induced SAPK activity, but not ERK2 phosphorylation, which remains elevated for up to 6 h. To examine directly the role of MKP-1 in the regulation of these protein kinases in vivo, we established a U937 cell line that conditionally expresses MKP-1 from the human metallothionein IIa promoter. Conditional expression of MKP-1 inhibited PMA-induced ERK2, SAPK, and p38 MAPK activity. By titrating the levels of MKP-1 expression from the human metallothionein IIa promoter, however, it was found that p38 MAPK and SAPK were much more sensitive to inhibition by MKP-1 than ERK2. This differential substrate specificity of MKP-1 can be functionally extended to nuclear transcriptional events in that PMA-induced c-Jun transcriptional activity was more sensitive to inhibition by MKP-1 than either Elk-1 or c-Myc. Conditional expression of MKP-1 also abolished the induction of endogenous MKP-1 protein expression in response to PMA treatment. This negative feedback regulatory mechanism is likely due to MKP-1-mediated inhibition of ERK2, as studies utilizing the MEK1/2 inhibitor PD98059 suggest that ERK2 activation is required for PMA-induced MKP-1 expression. These findings suggest that ERK2-mediated induction of MKP-1 may play an important role in preferentially attenuating signaling through the p38 MAPK and SAPK signal transduction pathways.

Mitogen-activated protein (MAP) 1 kinases play a key role in transducing various extracellular signals to the nucleus (1). The MAP kinases (MAPKs) consist of three major subgroups that include the ERK, SAPK/JNK, and p38 MAPK families (2). The ERKs, SAPKs, and p38 MAPK are activated by the reversible dual threonine and tyrosine phosphorylation of a conserved TEY, TPY, or TGY motif, respectively (1)(2)(3)(4)(5). Although distinct and selective activators of the MAPKs have been cloned and characterized (6 -9), less in known about the negative regulation of these kinases. The reversible nature of MAPK phosphorylation suggests that protein phosphatases play an important role in regulating MAPK activity. An expanding subfamily of dual specificity protein tyrosine phosphatases has been identified which is capable of dephosphorylating and inactivating various members of the MAPK family. This class of phosphatases is characterized by MKP-1 (also known as CL100, 3CH134, and Erp) (10 -13). MKP-1 expression is highly inducible in response to mitogenic and stress stimuli (10 -13). In this regard, the kinetics of serum-induced MKP-1 expression correlate with the inactivation of ERK2, suggesting a physiological role for MKP-1 in the attenuation of ERK2 activity (14). However, the rapid first phase inactivation of ERK2 after growth factor stimulation occurs prior to the induction of MKP-1 protein expression (15). ERK2 inactivation can also occur under conditions where MKP-1 protein synthesis is inhibited (15). Furthermore, the magnitude and duration of serum-induced ERK2 activity are unaltered in MKP-1-deficient mouse embryo fibroblasts (16). Thus, it is currently unclear whether MKP-1 is a physiological regulator of ERK2 in vivo.
Although initial reports indicated that MKP-1 selectively dephosphorylates ERK2 both in vivo and in vitro (14,17), recent studies suggest that MKP-1 exhibits a broad substrate specificity, being capable of inactivating both SAPK and p38 MAPK in transient transfection studies (5, 18 -22). PAC1 and MKP-2 have also been shown to inactivate multiple members of the MAPK family (21,23,24). Thus, certain dual specificity protein tyrosine phosphatases appear to exhibit a high degree of cross-reactivity with regard to substrate specificity. In this study we have examined the correlation between MKP-1 expression and the inhibition of ERK2, SAPK, and p38 MAPK in U937 human leukemic cells. To examine the in vivo substrate specificity of MKP-1 in detail we have established a U937 cell line that conditionally expresses MKP-1. We demonstrate that at sufficient levels of expression, MKP-1 is capable of inactivating all members of the MAPK family. However, titration of the levels of MKP-1 expression revealed that MKP-1 preferentially inactivates p38 MAPK and SAPK. This specificity can be extended to nuclear transcriptional events as MKP-1 preferentially inhibits PMA-induced c-Jun transcriptional activity when compared with both Elk-1 and c-Myc activity. Furthermore, we provide evidence that ERK2, and not p38 MAPK or SAPK, mediates MKP-1 induction in U937 cells. These findings suggest the existence of a cross-talk mechanism between the mitogen-and stress-induced signal transduction pathways, whereby ERK2-mediated induction of MKP-1 may play an important role in attenuating signaling through the p38 MAPK and SAPK pathways.
Cell Culture and Transfections-U937 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum (31). Transient transfections were performed using the DEAE-dextran procedure as described previously (32). Stable transfections were performed by electroporation (26). Two days post-transfection, cells were cultured in selective media containing 200 g/ml hygromycin B (Calbiochem). After 3 weeks of selection, resistant cell populations were pooled and cultured continuously in the presence of hygromycin B.
Western Blot Analysis and Immunoprecipitation-Whole cell extracts were resolved by SDS-polyacrylamide gel electrophoresis and electroblotted onto nitrocellulose. Endogenous ERK2 and MKP-1 expression were analyzed by immunoblotting with either an ERK2 monoclonal or MKP-1 polyclonal antibody as described previously (19). For analysis of inducible MKP-1(myc) protein expression from the hMTIIa promoter, U937 cells (5 ϫ 10 6 ) were metabolically labeled in the absence or presence of 0.5 M CdSO 4 for 4 h in methionine-free Dulbecco's modified Eagle's medium containing 10% dialyzed bovine calf serum and 0.1 mCi/ml Tran 35 S-label (NEN Life Science Products). Whole cell extracts were prepared as described previously (25,26,31) and MKP-1(myc) protein immunoprecipitated for 4 h using 9E10 antibody pre-coupled to protein A/G-Sepharose (Oncogene Science). Immunoprecipitates were washed extensively in cell lysis buffer, resolved by SDSpolyacrylamide gel electrophoresis, and visualized by fluorography.
Transcriptional Studies-U937 cells (12 ϫ 10 6 ) stably expressing either hMTIIa vector or hMTIIa-MKP-1(myc) were transiently transfected in batch with 10 g of the indicated Luc reporter plasmid. In the GAL4 chimeric activator transcriptional studies, cells were cotransfected with 10 g of the GAL4 chimeric activator and 10 g of a 5X-GAL4-Luc reporter, which consists of five GAL4 binding sites upstream of a luciferase reporter gene. Cells were split into 12-well plates and cultured overnight (16 -24 h). Cells were treated as indicated and cell extracts analyzed for luciferase activity.

RESULTS
PMA is a potent activator of both ERK2 and SAPK in U937 cells (19,31). In an attempt to determine whether MKP-1 may be a physiological regulator of these protein kinases, we compared the kinetics of PMA-induced MKP-1 expression with the inactivation of PMA-induced ERK2 and SAPK activity in U937 cells. PMA treatment caused a rapid induction of MKP-1 protein expression (Fig. 1, top panel). MKP-1 protein could be detected within 30 min, and maximal levels of expression were observed at 4 h of PMA treatment. PMA treatment was also found to induce a rapid and transient activation of SAPK ( Fig.  1, middle panel). Furthermore, the attenuation of SAPK activity coincided with the appearance of MKP-1 protein expression. In contrast to the rapid attenuation of PMA-induced SAPK activity, PMA treatment resulted in a prolonged phosphorylation of ERK2 as judged by its characteristic shift in electrophoretic mobility (Fig. 1, bottom panel). Although some reduction in ERK2 phosphorylation could be detected after 2 h of PMA treatment, ERK2 phosphorylation remained elevated for up to 6 h. Therefore, the kinetics of PMA-induced MKP-1 expression correlate more with the attenuation of SAPK activity than with the dephosphorylation of ERK2 in U937 cells.
To examine directly the role of MKP-1 in the regulation of SAPK and ERK2 in vivo, we established a U937 cell line conditionally expressing MKP-1 from the heavy metal-inducible hMTIIa promoter. U937 cells were transfected with either hMTIIa vector alone or hMTIIa-MKP-1 (containing a COOHterminal Myc epitope tag) and stable cell lines derived by selection in the presence of hygromycin B. No MKP-1 protein could be detected in untreated or CdSO 4 -treated hMTIIa vector-transfected control cells (Fig. 2). Although a low basal level of MKP-1 was observed in uninduced hMTIIa-MKP1 cells, MKP-1 expression was enhanced 40-fold after induction with CdSO 4 for 4 h (Fig. 2). The establishment of a U937 cell line conditionally expressing MKP-1 allowed us to examine the in vivo substrate specificity of MKP-1. As previous and preliminary studies had demonstrated that PMA was a potent activator of ERK2, SAPK, and p38 MAPK in U937 cells, we were able to examine whether conditional expression of MKP-1 functionally inhibited these kinases under identical cellular conditions. Cells were pretreated with 0.5 M CdSO 4 for 4 h prior to treatment with PMA for 30 min. Pretreatment of the hMTIIa vector-transfected U937 cell line with CdSO 4 had no effect on PMA-induced ERK2 phosphorylation or ERK2, SAPK, or p38 MAPK activity (Fig. 3, A, B, and C, respectively). Conditional expression of MKP-1 in the hMTIIa-MKP1 cell line, however, abolished PMA-induced ERK2 phosphorylation and activity (Fig. 3A). Conditional expression of MKP-1 was also found to inhibit PMA-induced SAPK and p38 MAPK activity (Fig. 3, B and C, respectively).
To examine more closely the in vivo substrate specificity of MKP-1, we attempted to vary the level of MKP-1 expression from the hMTIIa promoter. The activity of the hMTIIa promoter is highly sensitive to changes in heavy metal concentrations. To determine whether the magnitude of MKP-1 protein expression from the hMTIIa promoter could be regulated by varying the CdSO 4 concentration, MKP-1 was immunoprecipitated from the hMTIIa-MKP1 U937 cell line metabolically labeled for 4 h with [ 35 S]methionine in the absence or presence of increasing concentrations of CdSO 4 . Inducible MKP-1 protein expression was found to be proportional to the concentration of CdSO 4 utilized to induce the hMTIIa promoter (Fig. 4A). Thus, the sensitivities of ERK2, SAPK, and p38 MAPK to inhibition by various levels of MKP-1 expression were analyzed. Consistent with the results presented in Fig. 3, maximal induction of MKP-1 with 1 M CdSO 4 resulted in the inhibition of PMAinduced ERK2, SAPK, and p38 MAPK activity (Fig. 4B). However, titration of the level of MKP-1 expression revealed that p38 MAPK and SAPK were significantly more sensitive to inhibition by MKP-1 than ERK2 (Fig. 4B). In this regard, PMA-induced p38 MAPK activity was inhibited completely when little MKP-1 protein expression could be detected. These results suggest that MKP-1 may play a more selective role in attenuating signaling through the p38 MAPK and SAPK pathways than the ERK2 pathway in U937 cells.
SAPK mediates the phosphorylation of the c-Jun NH 2 -terminal transcriptional activation domain, which enhances c-Jun transcriptional activity (3,4,26,35). Because the conditional expression of MKP-1 abolished PMA-induced SAPK activity, we sought to determine whether this effect led to the inhibition of PMA-induced c-Jun transcriptional activity. To analyze c-Jun transcriptional activity, cells were transfected with a luciferase (Luc) reporter gene containing five upstream GAL4 DNA binding sites together with a GAL4-cJun(5-89) chimeric activator plasmid containing the c-Jun NH 2 -terminal transcriptional activation domain fused to the GAL4 DNA binding domain (19,26). PMA stimulated a 10 -15 fold-induction of GAL4-cJun transcriptional activity in the hMTIIa-MKP1 cell line (Fig. 5A). Preinduction of MKP-1 expression with CdSO 4 , however, abolished PMA-induced GAL4-cJun activity. c-Jun is ERK2 phosphorylation was determined as described for Fig. 1. Lower, ERK2 kinase activity was measured by an in-gel kinase assay using myelin basic protein as substrate. Panel B, SAPK was measured as described for Fig. 1. Panel C, FLAG-p38 MAPK activity was determined by immunocomplex kinase assay using anti-FLAG immunoprecipitates and GST-ATF2(1-109) as substrate. a major component of the AP-1 transcription factor which mediates the induction of various genes whose promoters contain TRE sites, including c-jun and c-fos. Thus, we examined the effect of MKP-1 on a Luc-reporter containing three TRE sites (3X-TRE-Luc), as well as Jun-Luc and Fos-Luc reporters, which contain the jun and fos promoters, respectively. PMAinduced TRE-Luc, Jun-Luc, and Fos-Luc activity was also inhibited by the conditional expression of MKP-1 (Fig. 5, B, C, and D, respectively). MKP-1 does not function as a general inhibitor of activated transcription as conditional expression of MKP-1 was found to have no effect on PMA-induced activation of a 4X-NFB-Luc reporter (Fig. 5E). Furthermore, CdSO 4 pretreatment of U937 cells stably expressing the hMTIIa vector alone had no effect on any of these PMA-induced transcriptional responses (data not shown).
The studies described above demonstrate that conditional expression of MKP-1 at levels that inhibit PMA-induced ERK2, SAPK, and p38 MAPK activity blocks a wide variety of PMAinduced transcriptional responses. To examine more closely whether the preferential inhibition of SAPK by MKP-1 results in the selective inhibition of specific PMA-induced transcriptional responses, we examined the effect of conditionally expressing various levels of MKP-1 on PMA-induced activation of several individual transcription factors. SAPK and ERK2 appear to be the sole mediators of c-Jun and c-Myc NH 2 -terminal phosphorylation and activation, respectively (3,4,7,8,36). Thus, we employed the GAL4-cJun(5-89) activator as an indicator of SAPK-mediated transcription and a GAL4-cMyc(7-101) activator as an indicator of ERK2-mediated transcription. The effect of MKP-1 on a GAL4-Elk-1(183-428) activator was also examined because of its coordinate regulation by both SAPK and ERK2 (37). Cells stably expressing hMTIIa-MKP1 were cotransfected with these GAL4 chimeric activator plasmids and a 5X-GAL4-Luc reporter. After induction of increasing levels of MKP-1 protein expression with various concentrations of CdSO 4 , cells were treated with PMA for 12 h and luciferase activity determined. PMA-induced GAL4-cJun transcriptional activity was found to be 10-fold more sensitive to inhibition by conditional expression of MKP-1 than GAL4-cMyc (IC 50 ϭ 12 nM and 127 nM CdSO 4 , respectively) (Fig. 6). In contrast, GAL4-Elk-1 exhibited an intermediate sensitivity to inhibition by MKP-1 (IC 50 ϭ 38 nM CdSO 4 ). These findings suggest that the differential substrate specificity of MKP-1 with regard to PMA-induced SAPK and ERK2 activity can be functionally extended to SAPK-and ERK2-mediated nuclear transcriptional events.
MKP-1 is an immediate-early gene product that is induced rapidly in response to both mitogenic and stress stimuli (10, 11, 13-15, 20, 38 -42). Although PMA selectively activates the ERK2 signal transduction pathway in most cell types (3-5), U937 cells provide a unique model in which PMA potently stimulates not only ERK2 activity but also SAPK and p38 MAPK activity (19, 26, 31, 36 and Figs. 3 and 4). Furthermore, both the SAPK and ERK2 signal transduction pathways have been implicated in the regulation of MKP-1 expression in response to various extracellular stimuli (20,38,39). We therefore examined the role of these signal transduction pathways in mediating PMA-induced MKP-1 expression in U937 cells. To determine the role of SAPK in mediating the induction of MKP-1 protein, we utilized UV radiation to activate selectively SAPK and not ERK2 activity in U937 cells. UV radiation (100 J/m 2 ) elicited a rapid transient activation of SAPK in U937 cells (Fig. 7A). The time course and fold activation of SAPK in response to UV radiation closely paralleled those observed in response to PMA treatment (compare with Fig. 1). UV radiation also stimulated p38 MAPK activity with a time course and potency similar to those observed for SAPK (data not shown). In contrast to PMA treatment (Fig. 7B, lanes 9 and 10), however, UV radiation had no effect on ERK2 phosphorylation (Fig.  7B, upper panel). Furthermore, although PMA induced a dra- matic induction of MKP-1 protein, none could be detected in response to UV radiation (Fig. 7B, lower panel). Prolonged exposure of these Western blots did not reveal activation of ERK2 or induction of MKP-1 in response to UV radiation (data not shown). These findings suggest that activation of the SAPK and p38 MAPK pathways is not sufficient to induce MKP-1 protein expression in U937 cells. These findings also indicate that inactivation of UV-induced SAPK and p38 MAPK does not involve induction of MKP-1 protein.
Although activation of p38 MAPK was not sufficient to induce MKP-1 expression in response to UV radiation, we examined directly the involvement of p38 MAPK activity in mediating PMA-induced MKP-1 expression by utilizing the specific p38 MAPK inhibitor SB203580 (43). Pretreatment of U937 cells with SB203580 was found to inhibit PMA-induced p38 MAPK activity in a dose-dependent manner, with 0.3 M SB203580 inhibiting p38 MAPK by 80% (Fig. 8, upper panel). Consistent with previous studies (43), SB203580 had no effect on PMA-induced ERK2 or SAPK activity in U937 cells (data not shown). Paradoxically, higher concentrations of SB203580 resulted in less inhibition of PMA-induced p38 MAPK activity (data not shown). Importantly, inhibiting PMA-induced p38 MAPK activity by 80% had no effect on PMA-induced MKP-1 expression (Fig. 8, lower panel). In aggregate, these studies suggest that activation of the SAPK and p38 MAPK pathway are neither required nor sufficient for induction of MKP-1 expression in U937 cells.
Many stimuli that induce MKP-1 expression, such as serum and mitogenic growth factors, potently activate the ERK2 pathway yet have little effect on SAPK and p38 MAPK activity. We therefore examined whether the ERK2 pathway was required for PMA-induced MKP-1 expression. PMA-induced activation of ERK2 is mediated by the sequential activation of Raf-1 and MEK1/2 (1,2). Thus the MEK1/2-specific inhibitor PD98059 was utilized to block PMA-induced ERK2 activity (44). Pretreatment of U937 cells with 100 M PD98059 resulted in a dramatic inhibition in PMA-induced ERK2 phosphorylation (Fig. 9, upper panel). Furthermore, this attenuation in PMAinduced ERK2 activation resulted in a near complete inhibition of PMA-induced MKP-1 expression (Fig. 9, lower panel). The residual induction of MKP-1 in the presence of PD98059 is likely because of the incomplete inhibition of PMA-induced ERK2 activity. These findings suggest that the ERK2 pathway predominantly mediates PMA-induced MKP-1 expression in U937 cells.
We have shown that conditional expression of MKP-1 is capable of inhibiting PMA-induced ERK2 activity and ERK2mediated gene transcription and that ERK2 mediates PMAinduced MKP-1 expression. It was thus of interest to determine whether conditional expression of MKP-1 blocked PMA-induced endogenous MKP-1 expression. Cells stably expressing hMTIIa-MKP1 were pretreated with increasing concentrations of CdSO 4 to inhibit preferentially the various PMA-induced protein kinase and transcriptional activities (see Figs. 4 and 6). Cells were then treated with PMA for 4 h, which resulted in maximal endogenous MKP-1 protein expression (see Fig. 1). Preinduction of hMTIIa-MKP-1 resulted in a dose-dependent inhibition of PMA-induced endogenous MKP-1 protein expression (Fig. 10). The efficacy of this inhibition was similar to the inhibition of PMA-induced ERK2 activity, although no inhibition of PMA-induced MKP-1 expression was observed at CdSO 4 concentrations that completely inhibited PMA-induced p38 MAPK activity (compare with Fig. 4). These findings suggest that MKP-1 regulates its own expression by a classical negative feedback regulatory mechanism. DISCUSSION MKP-1 was originally identified as a PMA-and mitogeninducible immediate-early gene product (11,13). MKP-1 was subsequently found to be a dual specificity phosphatase, exhibiting some selectivity for dephosphorylating and inactivating ERK2 (12,14). More recent studies utilizing transient transfection techniques, however, demonstrate that MKP-1 exhibits a broad substrate specificity, being capable of inactivating not only ERK2 but also SAPK and p38 MAPK (19 -22). In this FIG. 7. Activation of the SAPK pathway is not sufficient to induce MKP-1 protein expression in U937 cells. Panel A, parental U937 cells were exposed to UV radiation (100 J/m2) and cultured for the time periods indicated (lanes 2 and 3, min; lanes 4 -8, h). Whole cell extracts were analyzed for SAPK activity utilizing the GST-cJun(5-89) solid phase kinase assay (35). Panel B, upper, U937 cells were exposed to UV radiation as described above and ERK2 phosphorylation assessed by Western blot analysis. Lower, MKP-1 expression was determined by Western blot analysis. Parental U937 cells were also treated with PMA for 15 min and 4 h as positive controls for ERK2 phosphorylation and MKP-1 expression, respectively (lanes 9 and 10).  (13,17,38,45). An additional advantage of the U937 cell system involves the unusual ability of PMA to activate potently not only ERK2 but also the SAPK and p38 MAPK signal transduction pathways. Although the molecular mechanisms mediating PMA-induced SAPK and p38 MAPK activity in U937 cells have not been elucidated, this unique characteristic allowed us to examine the functional effects of MKP-1 on ERK2, SAPK, and p38 MAPK in response to a single extracellular stimulus.
In accordance with previous reports (11,13,40), PMA was found to induce MKP-1 protein expression in U937 cells rapidly. The time course of MKP-1 expression coincided with the rapid attenuation of PMA-induced SAPK activity, whereas ERK2 phosphorylation remained elevated for up to 6 h. This correlation led us to hypothesize that MKP-1 may inactivate SAPK preferentially with greater specificity than ERK2 in U937 cells. Although conditional expression of high levels of MKP-1 in U937 cells abolished PMA-induced ERK2, SAPK, and p38 MAPK activity, by titrating the levels of MKP-1 expressed from the hMTIIa-MKP1 expression vector we were able to demonstrate that MKP-1 inactivated p38 MAPK and SAPK preferentially compared with its effect on ERK2 activity. These findings not only substantiate previous studies demonstrating a broad substrate specificity for MKP-1 (5, 18, 19 -22) but also indicate that the selectivity of MKP-1 for p38 MAPK, SAPK, and ERK2 is highly dependent on the relative level of MKP-1 expression. These results appear to contradict a recent report that suggests that transiently transfected MKP-1 inactivates these kinases with identical substrate specificity (21). However, only two concentrations of MKP-1 cDNA were utilized to demonstrate substrate specificity, with the lowest amount analyzed inhibiting 50 -100% of the measured kinase activity in response to various extracellular stimuli (21). Therefore, this apparent discrepancy may merely be due to our more extensive analysis of this dose-dependent effect. Our findings are consistent with a model in which PMA-induced SAPK and p38 MAPK are inactivated rapidly when PMA-induced MKP-1 expression is submaximal (1-2 h), whereas ERK2 inactivation requires much higher levels of MKP-1 expression such as those observed after 4 -6 h of PMA treatment. However, the physiological relevance of this concentration effect is presently unclear. The studies described in this report were performed utilizing a MKP-1 expression vector encoding a COOH-terminal truncation (14,17). Unfortunately, all commercially available antibodies to MKP-1 are directed against the COOHterminal region of MKP-1, thus precluding a direct comparison of the levels of MKP-1 expressed from the hMTIIa promoter relative to that observed in response to PMA treatment.
Both mitogenic and stress stimuli have been reported to induce MKP-1 expression in various cell types (10, 11, 13-15, 20, 38 -42). Furthermore, both the SAPK and ERK2 signal transduction pathways have been implicated in the regulation of MKP-1 expression (20,39). Here we report that PMA, which activates both of these pathways potently, induces MKP-1 expression rapidly. The MEK1/2 inhibitor PD98059 dramatically reduced PMA-induced ERK2 activation and MKP-1 expression, suggesting that activation of the ERK2 signal transduction pathway is necessary for PMA-induced MKP-1 expression in U937 cells. Furthermore, the inability of UV radiation, which activates SAPK and p38 MAPK potently and selectively, to induce MKP-1 expression, suggests that the SAPK and p38 MAPK pathways do not play a significant role in mediating MKP-1 expression in U937 cells. These findings are in agreement with those of Brondello et al. (38), who have demonstrated that activation of SAPK by various stress stimuli does not induce MKP-1 protein expression, whereas selective activation of the Raf/MEK/ERK pathway was necessary and sufficient for MKP-1 expression in hamster fibroblasts. However, our results contradict previous reports that suggest that UVinduced SAPK activity is capable of mediating MKP-1 expression in HeLa and NIH3T3 cells (20,39). This apparent discrepancy may be because of the ability of UV radiation to activate ERK2 in a cell type-specific manner. In this regard, UV radiation has been shown to activate ERK2 in HeLa cells (47). However, we were unable to detect ERK2 activation in U937 cells in response to UV radiation (Fig. 7). Although higher doses of UV radiation may be capable of inducing ERK2 activation in U937 cells, our results indicate that doses that activate both SAPK and p38 MAPK potently have no effect on MKP-1 expression.
Although our data are consistent with a model in which PMA-induced MKP-1 inactivates SAPK and p38 MAPK preferentially, it is presently unclear whether MKP-1 is a physiological phosphatase for either of these kinases. In this regard, UV-induced SAPK is inactivated rapidly in the absence of MKP-1 induction in U937 cells, suggesting that additional dual specificity phosphatases are capable of mediating the attenuation of stress-induced SAPK activity in the absence of ERK2mediated MKP-1 expression. Recently, several additional MKP-1 family members have been identified which are capable of inactivating SAPK and/or p38 MAPK, including MKP-2 and PAC1 (21,23,38). Furthermore, the cytosolically localized dual specificity phosphatase M3/6 has been shown to exhibit an even greater selectivity for inactivating SAPK and p38 MAPK than that demonstrated here for MKP-1 (46). Whether these or other members of the MKP-1 family play a physiological role in attenuating signaling through the SAPK and p38 MAPK pathways remains to be determined. In aggregate, our findings are consistent with a model in which activation of the mitogenactivated signal transduction pathway and subsequent ERK2mediated MKP-1 expression may play an important role in preferentially attenuating signaling through the stress-acti-  1-9) or hMTIIa vector (lanes 10 -12) were pretreated for 4 h with the indicated concentration of CdSO 4 prior to treatment with PMA for 4 h. Cells were harvested and endogenous MKP-1 expression determined by Western blot analysis. vated protein kinases, p38 and SAPK. The existence of such a negative cross-talk mechanism between the mitogen-and stress-activated signal transduction pathways has significant physiological relevance based on the biological effects mediated by the ERK2 and p38/SAPK pathways. There is growing evidence to indicate that these pathways appear to play antagonistic roles in regulating either cellular growth, differentiation and/or survival (ERK2), or cell death (p38/SAPK) (48). ERK2mediated induction of MKP-1 may play an important role in selectively attenuating stress-induced signaling and directing cell differentiation, growth, or survival rather than cell death. Furthermore, the ability of MKP-1 to inhibit its own expression through a negative feedback mechanism reveals an additional level of regulation whereby MKP-1 not only attenuates signaling through the ERK, SAPK, and p38 MAPK pathways but may also play a role in resensitizing cells to subsequent extracellular stimuli by reducing its own expression and perhaps that of other dual specificity phosphatases.