Role of MEKK1 in Cell Survival and Activation of JNK and ERK Pathways Defined by Targeted Gene Disruption

See allHide authors and affiliations

Science  04 Dec 1998:
Vol. 282, Issue 5395, pp. 1911-1914
DOI: 10.1126/science.282.5395.1911


Targeted disruption of the gene encoding MEK kinase 1 (MEKK1), a mitogen-activated protein kinase (MAPK) kinase kinase, defined its function in the regulation of MAPK pathways and cell survival. MEKK1−/− embryonic stem cells from mice had lost or altered responses of the c-Jun amino-terminal kinase (JNK) to microtubule disruption and cold stress but activated JNK normally in response to heat shock, anisomycin, and ultraviolet irradiation. Activation of JNK was lost and that of extracellular signal-regulated protein kinase (ERK) was diminished in response to hyperosmolarity and serum factors in MEKK1−/− cells. Loss of MEKK1 expression resulted in a greater apoptotic response of cells to hyperosmolarity and microtubule disruption. When activated by specific stresses that alter cell shape and the cytoskeleton, MEKK1 signals to protect cells from apoptosis.

MEKK1 is a 196-kD protein serine-threonine kinase that has regulatory sequences for binding the small guanine nucleotide–binding proteins Ras (1) and Cdc42/Rac (2), an NH2-terminal bidentate 14-3-3 binding site (3), and a putative pleckstrin homology domain. When transfected into various cell types, MEKK1 activates the JNK and ERK pathways (4, 5). Expression of the catalytic domain of MEKK1 preferentially activates the JNK pathway (5). In contrast, expression of the complete enzyme effectively activates both the JNK and ERK pathways (6). MEKK1 is activated through the epidermal growth factor (EGF) receptor in PC12 (7), Cos (3), and T47D cells (8); the formyl-Met-Leu-Phe receptor in neutrophils (9); and antigen ligation of the high-affinity immunoglobulin E receptor, FcɛR1, in mast cells (10). Catalytically inactive inhibitory mutants of MEKK1 block receptor activation of both the JNK and ERK pathways (2), suggesting MEKK1 has the potential to regulate both MAPK pathways. MEKK1 is also a caspase-3 substrate (11). Cleavage of MEKK1 at Asp874 releases from cell membranes a 91-kD COOH-terminal kinase domain that amplifies caspase activation and induces apoptosis (11).

In mammalian cells, it is unprecedented for a MAPK kinase kinase to regulate two MAPK pathways in response to different extracellular stimuli. To define the specific stimuli that require the 196-kD MEKK1 for activation of specific MAPK pathways, we generated mouse embryonic stem (ES) cell lines lacking MEKK1 expression by homologous recombination (Fig. 1A) (12). Mutation of both MEKK1 alleles (MEKK1−/−) was established by selection in a high concentration of G418. Expression of the 196-kD MEKK1 protein was diminished in MEKK1+/− cells and absent in MEKK1−/− cells (Fig. 1B). The expression of Raf-1, B-Raf, ERK, JNK, and p38 proteins was unchanged in MEKK1−/− cells (13).

Figure 1

Gene targeting of MEKK1. (A) Genomic analysis of ES cells. Genomic DNA was isolated from MEKK1+/+, MEKK+/−, and MEKK1−/−cells, digested with Eco R1, and analyzed by Southern blotting with a 5′ flanking probe. (B) Absence of MEKK1 protein in MEKK1−/− cells. Cell lysates from MEKK1+/+, MEKK1+/−, and MEKK1−/− cells were resolved by SDS-PAGE, transferred to filters, and probed with an antibody to the COOH-terminus of MEKK1. The faint band (*) is a nonspecific band that is not related to MEKK1. Four additional antisera to different MEKK1 epitopes did not detect this band in MEKK1 immunoblots. Mr, molecular mass.

MEKK1 is a strong activator of the JNK pathway (2, 5). However, the stress response pathways that require MEKK1 are not defined. To define the role of MEKK1 in activation of the JNK pathway by different extracellular inputs, we exposed MEKK1−/−and MEKK1+/+ cells to several different stress stimuli, including microtubule disruption by nocodazole, cold shock, hyperosmolarity, ultraviolet (UV) irradiation, anisomycin, and heat shock (Fig. 2, A to D) (14). In addition, MEKK1−/− and MEKK1+/+ cells were deprived of serum for 2 hours and then stimulated with fetal bovine serum (FBS) (15%) (Fig. 2E) or lysophosphatidic acid (LPA) (Fig. 2F). The results revealed a distinct selectivity for participation of MEKK1 in activating JNK in response to various stimuli. MEKK1 was not required for JNK activation in response to UV irradiation, anisomycin, or heat shock (Fig. 2A). MEKK1 was required for JNK activation in response to microtubule disruption by nocodazole (Fig. 2B) and cold shock (Fig. 2C). MEKK1 was also required for JNK activation in response to modest hyperosmolarity achieved by the addition of sorbitol at concentrations up to 200 mM to the growth medium (Fig. 2D). Sorbitol concentrations above 200 mM resulted in the activation of the JNK pathway in both MEKK1−/− and MEKK+/+ cells. This finding indicates that two mechanisms exist in ES cells to respond to hyperosmolarity and that only one requires MEKK1. Activation of the JNK pathway by serum factors evident in MEKK1+/+ cells was virtually absent in MEKK1−/− cells (Fig. 2E). The serum response was mimicked by LPA (Fig. 2F), indicating that MEKK1 is required for JNK activation by the primary mitogen in serum.

Figure 2

Stimulation of JNK in MEKK1+/+ and MEKK1−/− cells. JNK activation was measured in response to (A) UV irradiation (100 J/m2), 42°C heat shock (HS) for 30 min, or anisomycin (AN) (50 ng/ml) for 30 min, (B) two concentrations of nocodazole for 1 hour, (C) cold stress induced by incubation at 18°C, (D) various concentrations of sorbitol for 30 min, (E) various times with 15% FBS, and (F) various times with 20 μM LPA. The results are for MEKK1+/+ cells (squares) and two independent MEKK1−/− clones (circles and diamonds). Shown is one experiment that is representative of three to seven experiments for the different clones and conditions.

Table 1 summarizes results from three to eight independent experiments with two MEKK1−/− clones. The results show that basal JNK activity is significantly diminished in MEKK1−/− cells. Furthermore, the stimulation of JNK activity in response to serum, LPA, nocodazole, low-dose sorbitol, and cold stress is markedly blunted. In contrast, the JNK stimulation in response to heat shock, anisomysin, and UV irradiation is similar in wild-type and two MEKK1−/− clones.

Table 1

Regulation of JNK, ERK, and p38 activity in wild-type and MEKK1−/− ES clones. JNK, ERK, and p38 activity data from three to eight experiments were pooled and statistically analyzed for each stimulus for wild-type (+/+) and two MEKK1−/−(−/−) ES cell clones. Values were determined by phosphorimager analysis, where basal activity was given a value of 1. Fold stimulation was then analyzed for each stimulus. Data are presented as the mean ± SEM. Where SEM is not shown, the data are the mean of two experiments. Incubation times for each stimulus were 10 and 30 min for ERK and p38, respectively. For JNK activation, incubation times varied: FBS, LPA, sorbitol, heat shock, and anisomycin, 30 min; nocodazole, 1 hour; cold shock, 2 hours; and UV, 1 hour after irradiation.

View this table:

Direct measurement of MEKK1 activity, assessed by its ability to phosphorylate recombinant MKK4 (also referred to as SEK1) (15), a MAPK kinase in the JNK pathway (16), confirmed that MEKK1 was activated in response to nocodazole, cold temperature, sorbitol, and serum (Fig. 3). MEKK1 also showed a shift to a slower mobility during SDS–polyacrylamide gel electrophoresis (SDS-PAGE) when isolated from MEKK1+/+ cells exposed to sorbitol. The gel shift is related to MEKK1 phosphorylation and activation (11). Thus, MEKK1 is activated and required for stimulation of the JNK pathway in response to microtubule disruption, mild hyperosmolarity, cold stress, or serum factors, including LPA. Activation of the JNK pathway in response to other stress stimuli, including UV irradiation, anisomycin, and heat shock, did not require MEKK1. Evidently, other MAPK kinase kinases must respond to these stress stimuli and activate the JNK pathway.

Figure 3

Activation of MEKK1 in MEKK1+/+ cells. Wild-type ES cells were (A) treated with 200 mM sorbitol for the indicated times, (B) treated with cold stress (20°C) for 1 and 2 hours, (C) deprived of FBS and leukemia inhibitory factor for 2 hours and then treated for 5 or 10 min with FBS (15%), or (D) incubated with nocodazole (0.05, 0.5, or 1 μg/ml) for 1 hour. MEKK1 activity was assayed in immunoprecipitates with recombinant catalytically inactive MKK4 as substrate. Fold activation over basal MEKK1 activity was determined by phosphorimaging and is shown below each lane. In (A), MEKK1 was resolved by SDS-PAGE, transferred to a filter, and immunoblotted (bottom panel).

The loss of JNK regulation by specific stimuli was demonstrated to directly result from the loss of MEKK1 expression. MEKK1−/− cells were stably transfected with the MEKK1 cDNA, and clones were screened for expression of the 196-kD enzyme. Two independent transfected MEKK1−/− clones that express MEKK1 driven by the cytomegalovirus promoter from the pCEP4 expression plasmid are shown in Fig. 4A. Higher levels of MEKK1 expression could not be obtained. We observed the cell cycle–dependent regulation of endogenous MEKK1 expression that is not mimicked when it is constitutively expressed from a plasmid (13). Nonetheless, add-back of MEKK1 at the expression levels achieved reconstituted the JNK response to nocodazole, LPA, and serum (Fig. 4B). In addition, cold stress inhibits JNK activity in MEKK1−/− cells, which is prevented by the add-back of MEKK1 expression, consistent with MEKK1 being required for JNK regulation during cold stress.

Figure 4

Restoration of the JNK and ERK response by expression of MEKK1 in MEKK1−/− cells. MEKK1−/− cells (−/−) were stably transfected with the cDNA for full-length MEKK1 in the expression plasmid pCEP4 and selected for resistance to hygromycin. (A) Clones expressing MEKK1 (AB1 and AB2) were identified by immunoblotting. (B) Expression of MEKK1 restored JNK activation in response to nocodazole (NOC; 0.5 μg/ml), 1 μM LPA, 15% FBS, and cold shock at 20°C (CS) relative to unstimulated cells having a basal (B) JNK activity. (C) Cells were serum starved for 2 hours and then challenged with 15% FBS for 10 min. ERK activity was measured after cell lysis and immunoprecipitation by phosphorylation of myelin basic protein. FBS-stimulated ERK activity in AB1 and AB2 was similar to that for the wild-type +/+ ES cells. MEKK1 add-back clones (AB1 and AB2) had a higher basal ERK activity relative to −/− clones. FBS-stimulated ERK activity is diminished in −/− cells. Basal activity, lanes 1, 3, 5, and 7; serum stimulation, lanes 2, 4, 6, and 8.

When overexpressed, MEKK1 activates the ERK and JNK pathways (2,4, 5). In vitro, MEKK1 phosphorylates and activates MKK1 (also referred to as MEK1), the MAPK kinase in the ERK pathway (4). Furthermore, catalytically inactive MEKK1 inhibits stimulation of ERK in response to EGF (2). ERK activity was modest in ES cells (17). Two-hour serum deprivation was observed to significantly diminish basal ERK activity in MEKK1−/− cells but not in wild-type ES cells (Table 1). ERK stimulation by sorbitol, serum, and LPA but not phorbol ester was also diminished in MEKK1−/− clones. These results, statistically analyzed for multiple independent experiments, show that MEKK1 contributes to ERK activation. However, the contribution of MEKK1 to ERK activation in response to different stimuli is only partial, because of the overlapping activation of Raf-1 and B-Raf, which also activate ERK. Expression of MEKK1 by stable transfection of the MEKK1−/− ES cells restored ERK activation in response to serum factors (Fig. 4C). Basal and serum-stimulated ERK activity was elevated in the add-back clones relative to MEKK1−/−cells, demonstrating that MEKK1 contributes to regulation of the ERK pathway in addition to the JNK pathway. Activation of p38 in response to hyperosmolarity was identical in MEKK−/− and MEKK+/+ cells (Table 1). MEKK1 does not appear to regulate p38.

Our results indicate that MEKK1 is a MAPK kinase kinase that selectively regulates two different MAPK modules. InSaccharomyces cerevisiae, the MAPK kinase kinase STE11 participates in MAPK pathways functioning in pheromone-regulated mating, pseudohyphal or invasive growth, and hyperosmolarity (18). The MAPK kinase for mating and pseudohyphal or invasive growth is STE7, whereas the MAPK differs: Fus 3 for mating and KSS1 for pseudohyphal or invasive growth. In the hyperosmotic response pathway, the MAPK kinase and MAPK are Pbs2 and Hog1, respectively (18, 19). We found that, like STE11, MEKK1 responds to different stimuli for the selective regulation of at least two MAPK pathways.

Nocodazole treatment of cells did not significantly activate ERK (13). Thus, MEKK1 activation in response to microtubule disruption selectively stimulates JNK. Nocodazole, cold stress, and mild hyperosmolarity all require MEKK1 for JNK activation. The common feature of each of these stresses is a cell shape change, suggesting that MEKK1 activates JNK in response to sensors that detect shape changes. MEKK1−/− cells have a diminished capacity to survive stresses including mild hyperosmolarity and nocodazole-induced disruption of microtubules (Fig. 5). Thus, MEKK1 is required for survival during environmental challenges that stress cells. In normal medium, the growth of MEKK1−/− and wild-type ES cells is similar. Treatment of ES cells with sorbitol inhibited the growth of MEKK1−/−cells relative to wild-type ES cells (Fig. 5A). In addition to sorbitol-induced growth arrest, the number of apoptotic cells was substantially greater for MEKK1−/− compared with wild-type ES cells (Fig. 5B). MEKK1−/− cells were also substantially more sensitive to nocodazole-induced apoptosis (Fig. 5, C and D). MEKK1 expression in transfected MEKK1−/− clones restored the survival response to that of wild-type ES cells just as it reconstituted JNK activation in response to nocodazole treatment.

Figure 5

MEKK1−/− cells have an increased apoptotic response to stress. (A andB) MEKK1+/+ and MEKK1−/− cells were grown in the presence of 200 mM sorbitol. Cell numbers were determined on days 1 to 5. The growth of −/− cells is substantially less than that of +/+ cells in the presence of sorbitol. Contributing to the decreased cell number for −/− clones is an increased apoptotic rate. (C and D) Nocodazole (NOC) treatment of ES cells disrupts microtubules and induces apoptosis. MEKK1−/− cells are more sensitive to nocodazole as measured by percentage of apoptotic cells 16 hours after exposure to drug. The add-back of MEKK1 expression (AB1 and AB2) rescued the −/− ES cells so that their sensitivity to nocodazole-induced apoptosis was similar to +/+ cells. Shown are two independent experiments where values represent the mean of triplicate samples ± SEM. Apoptosis was measured by acridine orange–ethidium bromide staining and also verified by terminal deoxytransferase analysis of DNA fragmentation for sorbitol treatment (13).

The antiapoptotic function of MEKK1 activation has been defined by its targeted disruption. The findings define the ability of MEKK1 to enhance the survival of cells after a stress response. MEKK1 promotion of cell survival is opposite to its described proapoptotic function initiated when it is cleaved by caspases (11). Caspase cleavage of MEKK1 releases a 91-kD activated kinase domain that further amplifies caspase activity committing cells to apoptosis; activated full-length 196-kD MEKK1 activates neither caspases nor apoptosis (20). Thus, there is a dual role for MEKK1 controlled by its cleavage. Caspases act as switches to convert the MEKK1 survival signal to a proapoptotic response.

  • * To whom correspondence should be addressed. E-mail: johnsong{at}


View Abstract

Stay Connected to Science

Navigate This Article