A Novel Regulatory Mechanism in the Mitogen-activated Protein (MAP) Kinase Cascade

Mitogen-activated protein kinase (MAPK) kinase (MAPKK, also known as MEK), a direct activator for MAPK/extracellular signal-regulated kinase, localizes to the cytoplasm excluded from the nucleus during signal transmission. This nuclear exclusion of MAPKK is directed by its nuclear export signal (NES), but its physiological significance has been unknown. We have found that disruption of the NES dramatically potentiates the ability of constitutively active MAPKK to induce morphological changes and malignant transformation of fibroblastic cells. Readdition of the NES sequence reversed the effects induced by the NES disruption. Moreover, we observed that a dramatic increase of activated MAPK in the nucleus was induced by the NES-disrupted MAPKK and that coexpression of MAPK phosphatase-1 (CL-100) or a kinase negative form of MAPK counteracted the phenotypes induced by the NES-disrupted MAPKK, indicating the crucial role of MAPK in the responses. These findings reveal a novel regulatory role of the NES of MAPKK that may be essential for proper signal transductions.

(reviewed in Ref. 13). The activated Ras recruits Raf-1 to the plasma membrane where the activation of Raf-1 occurs (14 -20). Then the activated Raf-1, as MAPKK kinase, phosphorylates and activates MAPKK, which in turn activates MAPK in the cytoplasm. These three kinases undergo multiple phosphorylation reactions in the cytoplasm, whereas the terminal component of the three kinases, MAPK, is translocated from the cytoplasm to the nucleus (21)(22)(23) where MAPK phosphorylates several nuclear targets such as Elk-1, Myc, and Myb (reviewed in Refs. 24 and 25). Thus, the MAPK cascade constitutes a functional unit that links cytoplasmic signaling events to nuclear events.
MAPKK stays in the cytoplasm during signal transmission (23,26,27). We have previously reported that MAPKK has in its near N-terminal region a short amino acid sequence that acts as a nuclear export signal (NES) and that this NES directs cytoplamic localization or nuclear exclusion of MAPKK (28). The concept of NES initially emerged from studies characterizing two proteins, human immunodeficiency virus Rev protein and PKI, an inhibitor of protein kinase A (29,30). The identified NES sequences are rich in hydrophobic residues, which are critical for the NES activity (29 -32). Although we proposed that the NES of MAPKK functions as a determinant for cytoplasmic localization of MAPKK (28), the biological role of the NES-directed cytoplasmic localization of MAPKK in signal transduction remains unclear.
Vertebrate MAPKK is activated by phosphorylation of two serine residues, Ser 218 and Ser 222 (33)(34)(35)(36)(37). Constitutively active mutants of MAPKK were successfully generated by substituting acidic amino acids for these two serine residues (33)(34)(35)(36)38), and deletion of the region near the N terminus of MAPKK outside the kinase catalytic core was reported to enhance the kinase activity of the mutant MAPKK (38). These constitutively active mutants of MAPKK were shown to be able to induce malignant cell transformation of fibroblasts (38,39), PC12 cell differentiation (39), and mesoderm formation in Xenopus (40 -42). In this study, we produced a number of NESdisrupted mutants of MAPKK by substituting alanines for critical leucine residues in the NES sequence or by deleting the NES sequence. These NES-disrupted MAPKKs became present both in the nucleus and in the cytoplasm. We have also found that the NES-disrupted constitutively active MAPKKs induce much stronger responses than active MAPKKs containing an intact NES, which are present exclusively in the cytoplasm, in cell shape changes of fibroblastic cells and malignant transformation of NIH 3T3 cells. Furthermore, readdition of the NES sequence to the NES-disrupted active MAPKKs makes their subcellular localization exclusively cytoplasmic and reverses the exaggerated effects caused by the NES disruption. Therefore, we suggest that the NES of MAPKK is a regulatory domain that suppresses abnormal cellular responses such as malignant transformation. In other words, the NES-mediated nuclear exclusion of MAPKK is essential for regulated signal transductions.

MATERIALS AND METHODS
Cell Staining-Cells were fixed and permeabilized as described previously (28). To determine subcellular localization of the HA-tagged proteins, cells were stained with anti-HA monoclonal antibody (12CA5) followed by a rhodamine-conjugated goat anti-mouse antibody (Cappel). Fluorescence was observed in an Axiophot with a 63 X plan-Neofl. (1.25 NA) objective.
The supernatant was subjected to immunoprecipitation with antibody to HA (12CA5). A portion of the supernatant was mixed with 3 l of the antibody and 20 l of protein A-Sepharose (Pharmacia Biotech, Inc.). 2 h after incubation at 4°C, the immunocomplex was washed three times with Tris-buffered saline containing 0.05% Tween 20 and once with kinase buffer (20 mM Tris-Cl, pH 7.5, 100 M ATP 10 mM MgCl 2 ). Immune complex kinase assays were performed in 15 l of kinase buffer containing 2 Ci of [␥-32 P]ATP and 3 g of GST-KN-MAPK (a recombinant kinase-negative Xenopus MAPK (K57D, glutathione S-transferase fusion protein)) as a substrate at 30°C for 15 min. The reaction was stopped by the addition of the SDS sample buffer and boiling. After SDS-polyacrylamide gel electrophoresis, 32 P incorporation into GST-KN-MAPK was quantified with an Image Analyzer (Fujix BAS2000).
Transfections-COS-7 cells and Swiss 3T3 cells were maintained in DMEM supplemented with 10% fetal calf serum and antibiotics. PC12 cells were cultured in DMEM supplemented with 10% fetal calf serum, 5% heat inactivated horse serum, and 0.35% glucose and antibiotics. Transfection with constructs encoding MAPKK mutants or MAPK was carried out by the LipofectAMINE method as recommended by the manufacturer (Life Technologies, Inc.). Briefly, cells were seeded in 6-well plates at a density of 2 ϫ 10 5 cells/well 16 h before transfection. The plasmids were mixed with LipofectAMINE (Life Technologies, Inc.) in serum-free Opti-MEM (Life Technologies, Inc.), and the cells were transfected. The total amount of DNA in transfection was 3 g/well. 5 h after transfection, the cells were incubated in DMEM supplemented with fetal calf serum.
Transformation Assay-NIH 3T3 cells were transiently transfected by lipofection with the indicated plasmids. For focus forming assay, cells were incubated in DMEM supplemented with 7% calf serum for the days indicated by the figure legends.

Production and Expression of Constitutively Active MAPKKs with an Intact NES or a Disrupted NES-
We made various mutants of MAPKK in which serine residues of activation phosphorylation sites (Ser 218 and Ser 222 ) were replaced by acidic amino acids (33)(34)(35)(36)(37)(38)(39), glutamic acid or aspartic acid, to produce constitutively active mutants (S222E is SE, S218E/ S222E is SESE, and S218D/S222E is SDSE in Fig. 1A) and critical leucine residues (Leu 33 and Leu 37 ) in the NES sequence (28) (Leu 33 -Leu 42 ) were replaced by alanines to produce the NES-disrupted MAPKK (LA mutants in Fig. 1A). We also generated a constitutively active MAPKK that lacked the region containing the NES (deletion of residues 32-51) (⌬-SESE in Fig. 1A) (38). We expressed each of these six mutant MAPKKs or wild type MAPKK as a HA epitope-tagged protein in COS-7 cells and determined the kinase activity. LA-MAPKK had a low basal kinase activity, which was comparable with or slightly higher than that of WT MAPKK (Fig. 1B). LA-SE MAPKK, LA-SESE MAPKK, LA-SDSE MAPKK, SDSE MAPKK, and ⌬-SESE MAPKK had about 10-, 20-, 50-, 35-, and 30-fold higher kinase activity than WT MAPKK, respectively (Fig. 1B). The effect of expression of these mutant MAPKKs on MAPK activity was examined by a cotransfection assay in COS-7 cells. WT MAPKK and LA MAPKK induced little or no activation of MAPK (Fig. 1C). Expression of LA-SE MAPKK, LA-SESE MAPKK, LA-SDSE MAPKK, SDSE MAPKK, and ⌬-SESE MAPKK resulted in activation of about one-fourth, one-third, four-fifths, two-thirds, and one-half the population of MAPK, respectively (Fig. 1C). Thus, the ability of these mutant MAPKKs to activate MAPK in cells roughly correlated with their kinase activity in vitro. Similar results were obtained in rat 3Y1 cells, NIH 3T3 cells, and PC12 cells (data not shown). When these mutant MAPKKs were cotransfected with other members of the MAPK superfamily, JNK/SAPK or p38, the kinase activity of the expressed JNK/SAPK or p38 was not activated, indicating that the mutant MAPKKs specifically act on classical MAPK (extracellular signal-regulated kinase) (data not shown).
Constitutively Active MAPKKs with a Disrupted NES Induce a Drastic Change in Cell Morphology and Cause Cell Transfor-mation Efficiently-Then we expressed these various MAPKKs (HA-tagged) in rat fibroblastic 3Y1 cells and examined their subcellular distribution. WT MAPKK and SDSE MAPKK localized to the cytoplasm and were excluded from the nucleus (Fig. 2, WT and SDSE), whereas LA-SE MAPKK, LA-SESE MAPKK, LA-SDSE MAPKK, and ⌬-SESE MAPKK were all present in both the nucleus and the cytoplasm (Fig. 2). Interestingly, we found that expression of LA-SESE MAPKK, LA-SDSE MAPKK, or ⌬-SESE MAPKK resulted in a drastic change in morphology of the cell (Fig. 2). Unlike normal flattened shape of the fibroblastic cell, these cells became shrunken and rounded, and some thin protrusions in the cell periphery remained. Microscopic observations showed that these deformed cells tend to move actively and ride on other cells. 2 In contrast, expression of SDSE MAPKK, which has a higher kinase activity than LA-SESE MAPKK, did not induce drastic changes in cell morphology: the cells remained flattened and did not lie on other cells, like normal cells (Fig. 2). These results suggest that constitutively active MAPKKs with a disrupted NES, which are present in both the nucleus and the cytoplasm, are able to induce the drastic morphological change, whereas constitutively active MAPKK with an intact NES that is present exclusively in the cytoplasm is unable to induce such a drastic effect. We found further in the experiments with PC12 cells that expression of ⌬-SESE MAPKK or LA-SDSE MAPKK induced neurite outgrowth much more potently than expression of SDSE MAPKK in the absence of nerve growth factor (data not shown). Thus, subcellular localization of active MAPKKs should be important for their ability to induce PC12 cell differentiation.
It has been shown that constitutively active MAPKKs can induce malignant transformation of NIH 3T3 cells (38,39). Then we compared the ability of the different MAPKKs to induce NIH 3T3 cell transformation in a focus forming assay. Expression of LA-SDSE MAPKK induced many foci, comparable with that of activated Ras. Also, LA-SDSE MAPKK and ⌬-SESE MAPKK induced a number of foci reproducibly. In contrast, SDSE MAPKK that has an intact NES did not pro-duce any foci under the conditions (Fig. 3). Immunoblotting confirmed that each of different MAPKKs was expressed to almost the same extent (data not shown). These results indicate that the constitutively active MAPKK, if present in both the cytoplasm and the nucleus due to the NES deficiency (disruption or deletion of NES), is able to cause cell transformation efficiently and induce a large number of foci. To test whether the NES disruption alone might induce some abnormal response, NIH 3T3 cells were transfected with either WT MAPKK or LA MAPKK and cultured in the presence of 15% fetal calf serum for 2 weeks. The medium was refreshed every other day. Remarkably, cells expressing LA MAPKK produced a number of foci, whereas cells expressing WT MAPKK formed no foci (data not shown). Thus, not only constitutive activation of the kinase activity of MAPKK but also its presence in the nucleus induced by the disruption of NES appears to be crucial for enhancing the efficiency of transformation.

Effect of Readdition of the NES Sequence to the NES-disrupted MAPKKs on Their Ability to Induce Morphological
Changes and Cell Transformation-To further confirm that the marked potentiation of the biological responses by the NES disruption of MAPKK results from the change in the subcellular distribution of MAPKK and not from any side effects caused by the mutations in the NES portion of MAPKK, we made MAPKK constructs in which the NES sequence of MAPKK (residues 32-44) was added to the N terminus of the NESdisrupted active MAPKKs. Thus, NES-LA-SDSE MAPKK and NES-⌬-SESE MAPKK constructs were newly produced (Fig.  4A). NES-⌬-SESE MAPKK and ⌬-SESE MAPKK had almost the same kinase activity, and NES-LA-SDSE MAPKK and LA-SDSE MAPKK also had almost the same kinase activity (Fig. 4B). A cotransfection assay with MAPK also showed that addition of NES did not change the ability of ⌬-SESE MAPKK or LA-SDSE MAPKK to activate MAPK in cells (Fig. 4C). To test the ability of these MAPKKs to induce the change in cell morphology, we injected the nuclei of rat 3Y1 cells with the plasmids expressing these MAPKKs. ⌬-SESE MAPKK and LA-SDSE MAPKK were present in both the nucleus and the cytoplasm, whereas NES-⌬-SESE MAPKK and NES-LA-SDSE MAPKK were completely cytoplasmic, excluded from the nu-2 I. Gotoh, M. Fukuda and E. Nishida, unpublished data.

FIG. 1. Characterization of constitutively active MAPKK mutants with an intact NES or a disrupted NES.
A, schematic representation of the MAPKK derivatives used in this study. To generate activated mutants, serine 218 or 222 or both were replaced by glutamic acid (E) or aspartic acid (D). In addition, two critical leucines in NES sequence of MAPKK were replaced by alanines to disrupt NES of MAPKK (LA). ⌬-SESE is a constitutively active MAPKK that is made by the combination of both deletion of residues 32-51 and substitutions of phosphorylation sites (SESE). B, kinase activity of mutant MAPKKs expressed in COS-7 cells. The expressed HA-tagged MAPKKs were immunoprecipitated with monoclonal antibody 12CA5. Immunoprecipitates were subjected to the kinase assay. The data points represent mean Ϯ standard error of mean for three separate experiments. C, effect of expression of mutant MAPKKs on MAPK activity in COS-7 cells. Cells lysates were prepared from COS-7 cells transfected with the indicated mutant MAPKK constructs together with the Xenopus MAPK construct. They were subjected to immunoblotting using an anti-Xenopus MAPK antibody. The experiment was repeated three times with similar results. cleus as expected (Fig. 4D). As described before, the former two MAPKKs induced a drastic change in cell morphology (Fig. 4D), but the latter two MAPKKs, the NES-readded mutants, did not induce marked morphological changes; the cells were spread and flattened (Fig. 4D). Then to compare the ability of these MAPKKs to transform cells, NIH 3T3 cells were transfected with each of these MAPKKs. The cells expressing the NESdisrupted active MAPKKs developed a large number of foci, but the cells expressing the NES-readded MAPKKs did not form foci under those conditions (Fig. 4E). These results taken together indicate clearly that the effects induced by the NES disruption result from the change in subcellular localization of MAPKK.

Requirement of MAPK Activation for Morphological Changes Induced by the NES-disrupted Active
MAPKKs-To address the mechanisms by which the presence of active MAPKK molecules in the nucleus potentiates dramatically these abnormal cellular responses, we expressed the mutant MAPKKs together with MAPK in COS cells and examined subcellular localization of activated MAPK. The cell staining with anti-active MAPK antibody showed that a dramatic increase of activated MAPK in the nucleus was induced by the NES-disrupted active MAPKKs, ⌬-SESE MAPKK and LA-SDSE MAPKK, whereas activated MAPK was present largely in the cytoplasm in cells expressing the NES-containing active MAPKK, SDSE-MAPKK (Fig. 5). Then we examined the role of MAPK in the responses induced by the NES-disrupted active MAPKKs. Swiss 3T3 cells were transfected with HA-tagged mutant MAPKKs with or without MAPK-specific phosphatase CL100 (46 -48). Expression of ⌬-SESE MAPKK alone or LA-SDSE MAPKK alone in Swiss 3T3 cells resulted in a drastic change in cell morphology (Fig. 6A), as in the case of rat 3Y1 cells (see Fig. 2). The cells expressing these MAPKKs moved actively and tended to lie on other cells, and stress fibers in them became diminished (Fig.   FIG. 2. NES-disrupted constitutively active MAPKKs induce morphological changes. Various mutant MAPKK (HA-tagged) constructs (150 g/ml) were injected into the nuclei of rat 3Y1 cells. Cells were stained with an anti-HA monoclonal antibody. Experiments were performed at least three times, and the results of a representative experiment are shown.

FIG. 3. Transformation of fibroblastic cells by constitutively
active MAPKKs with a disrupted NES. NIH 3T3 cells were transiently transfected with constructs encoding mutant MAPKKs or with a Ras V12 construct and incubated in 7% calf serum-DMEM. Plasmid concentrations were 5 g/10-cm dish. 3 weeks after transfection, foci on each dish were visualized and photographed with phase contrast microscopy (left panels). The appearance of transformed foci is clearly shown at higher resolution (right panels). Four further experiments showed similar results. 6A, ACTIN). Co-expression of CL100 almost completely counteracted these phenotypes induced by the constitutively active MAPKKs with the disrupted NES (Fig. 6A). The cells became flattened and did not lie on other cells. Stress fibers also became normal. These results suggest that the mutant MAPKKinduced change in cell morphology requires activation of MAPK. Next, NIH 3T3 cells were transfected with a kinasenegative form of MAPK (KN-MAPK) together with the mutant MAPKKs. Expression of LA-SDSE MAPKK alone induced a large number of foci, but co-expression of KN-MAPK greatly inhibited the focus formation (Fig. 6B). This confirmed the requirement of MAPK in the phenotypes induced by the mutant MAPKKs. DISCUSSION We have recently shown that MAPKK contains in its near N-terminal region a leucine-rich NES sequence that directs cytoplasmic localization, i.e. nuclear exclusion, of MAPKK (28). This is consistent with previous observations that MAPKK stays in the cytoplasm during signal transmission and thus may account for permanent cytoplasmic localization of MAPKK in cells. In fact, when the NES sequence is dis-  Fig. 1C, and the cell lysates were immunoblotted with an anti-Xenopus MAPK antibody. Similar results were obtained in three separate experiments. D, the nuclei of rat 3Y1 cells were injected with each of these mutant MAPKK constructs (150 g/ml). The cells were stained with 12CA5 as described in Fig. 2. E, NIH 3T3 cells were transiently transfected with the indicated mutant MAPKK constructs and cultured for 21 days in 7% calf serum-DMEM as described in the legend to Fig. 3. The fields were visualized by phase contrast microscopy with a low power lens (left panels) or a high power lens (right panels). Similar results were obtained in three independent experiments in D and E. rupted by substituting alanines for leucines that are important for the NES activity, cytoplasmic localization is broken and MAPKK becomes present in both the nucleus and the cytoplasm (Ref. 28 and this study). It remained unclear whether the NES-directed nuclear exclusion of MAPKK has some physiological significance. In this study, we have addressed this question by expressing a number of constitutively active MAPKKs with either a disrupted NES or an intact NES in various fibroblastic cells. In all cases the MAPKKs having an intact NES are excluded from the nucleus, and the MAPKKs with a disrupted NES are present in both the cytoplasm and the nucleus. The active MAPKKs with a disrupted NES induce morphological changes of fibroblastic cells and NIH 3T3 cell transformation (focus formation) much more strongly than the active MAPKKs with an intact NES. Importantly, this marked potentiation of the biological responses by the NES disruption is reversed by readdition of the NES sequence to the NES-disrupted MAPKKs. These NES-readded mutant MAPKKs are present in the cytoplasm, excluded from the nucleus, and have almost the same kinase activity as the original NES-disrupted ones. Therefore, we can conclude that the NES disruption-induced effects result from the change in the subcellular distribution of MAPKK.
The results presented here suggest that the NES-directed cytoplasmic localization or nuclear exclusion of MAPKK is crucial for regulated normal cell proliferation. In other words, the NES of MAPKK works as a suppressor of unregulated cell proliferation. The NES in MAPKK may thus be a novel regulatory domain that determines cytoplasmic localization of MAPKK and has an anti-oncogenic or anti-proliferative activity. Consistent with this hypothesis, expression of the NESdisrupted MAPKK whose kinase activity is not constitutively active but not expression of an intact NES-containing WT MAPKK results in focus formation in NIH 3T3 cells under the conditions that promote cell proliferation. Although we cannot exclude the possibility that the process of nuclear export of MAPKK has some role, the final location of MAPKK in the cytoplasm may have a primary importance.
Previous experiments from several laboratories (38,39) showed that expression of constitutively active MAPKKs with an intact NES is able to induce NIH 3T3 cell transformation, an observation apparently contradictory to the present results. This apparent discrepancy presumably results from the difference in the assay system. We utilized a transient transfection assay, but they produced cell lines stably expressing mutant MAPKKs to enhance the sensitivity of detection of the transforming activity. In the stable transfection assays also, it was observed that the NES-deleted MAPKK (⌬-S222D MAPKK) had a more potent ability to transform cells than the NEScontaining MAPKK, although Mansour et al. attributed the cause to the enhanced kinase activity of ⌬-S222D MAPKK (38). Our present data that the NES-readded mutant MAPKKs with the disrupted NES have the same kinase activity as the original mutant MAPKKs with the disrupted NES but have much a reduced ability to transform cells suggest that the change in subcellular localization is a primary cause of the difference. Anyway, LA-SDSE MAPKK that has a disrupted NES and a highest kinase activity is the most potent to induce the MAPK cascade-mediated cellular responses and therefore may be useful for other investigations.
In this study we have shown that when the NES-disrupted active MAPKK plasmids are injected into the nuclei of the fibroblastic cells, the plasmid-injected cells undergo drastic changes in cell morphology and the stress fibers become diminished, but surrounding uninjected cells do not undergo marked morphological changes. This result may suggest that the MAPK cascade-dependent morphological changes do not necessarily involve an autocrine mechanism. But this does not rule out the possibility of an intracellular autocrine mechanism in which newly synthesized ligands bind to the receptors within the cell. It remained answered whether the active MAPKK-mediated changes require new transcription or new protein synthesis. Also, the role of low molecular weight GTPases such as Ras, Rho, and Rac in this process should be elucidated.
Important issues are molecular mechanisms by which the presence of active MAPKK molecules in the nucleus dramatically potentiates these cellular responses. We have recently proposed that MAPKK can act as a cytoplasmic anchor for MAPK in unstimulated cells (45). Thus, nuclear exclusion of MAPK may also be directed by the NES of MAPKK. Then if the NES of MAPKK is disrupted, MAPKK-mediated nuclear exclusion of MAPK is broken and MAPK becomes nuclear. In fact, we revealed here that activated MAPK becomes nuclear in cells expressing the NES disrupted active MAPKKs (Fig. 5). In addition, we have shown that coexpression of MAPK phosphatase-1 (CL-100), which acts in the nucleus, cancels the effects induced by the mutant MAPKKs. Moreover, MAPK is the only known substrate for MAPKK, and it has been shown that other members of the MAPK superfamily are not activated by the mutant MAPKK in cells. Thus, we may propose that all the effects induced by the NES-disrupted active MAPKKs could be through activation of MAPK in the nucleus.
In summary, our present results reveal a hitherto unrecognized function of NES of MAPKK. The NES-directed nuclear exclusion of MAPKK may have a role in preventing the cells from undergoing abnormal responses such as malignant cell transformation. The NES might regulate subcellular distribution of MAPK as well. Through these activities, the NES of MAPKK may assure proper signal transductions mediated by the MAPKK/MAPK cascade. This study thus emphasizes the importance of subcellular localization of signaling molecules in the regulation of cellular responses.