INTRODUCTION

The role of heterotrimeric guanine nucleotide binding proteins (G-proteins)¹ in cell differentiation and development has been shown in several systems (1-4). Adipogenesis of NIH 3T3-L1 cells in response to inducers such as insulin or dexamethasone plus methylisobutylxanthine is accompanied by a sharp decline in G_s subunit (5). Both suppression of G_s by antisense oligodeoxynucleotides and overexpression of the constitutively active mutant form of G_i2 promote adipogenesis in the absence of classical inducers (5). Differentiation of F9 embryonic stem cells to endodermal cells provokes a sharp reduction of G_i2 (2), a de-repression of phospholipase C, and activation of protein kinase C and mitogen-activated protein (MAP) kinase, especially via ERK1,2 (6). The expression of constitutively active forms of G_q or G₁₆ can induce neuronal differentiation of PC 12 cells (7), via activation of c-Jun amino-terminal kinase but not via ERK1,2.

The extracellular signal-regulated kinase (ERK), stress-activated protein kinase (SAPK/JNK), and mammalian homolog of the yeast-osmosensing ERK HOG1 (p38 MAPK) are conserved members of a MAP kinase cascade for regulation of targets such as transcription factors in response to growth factors or environmental stresses, such as ultraviolet light, and protein synthesis inhibitors (8-16). ERK appears to play a major role in provoking cell proliferation and differentiation (17-19), whereas JNK mediates stress responses and some forms of apoptosis (20-23). JNK has been implicated also in the induction of differentiation (7, 24) and oncogenesis (25). The ability of overexpression of c-Jun, which is a target for JNK, to stimulate differentiation of P19 embryonal carcinoma cells to a mixed endoderm/mesoderm population supports a role of JNK in differentiation (26).

P19 embryonal carcinoma (P19) cells have been used as a model system for murine pre-implantation development (27, 28). These pluripotent cells have the ability to differentiate into derivatives of three germ layers, endoderm, mesoderm, and ectoderm, upon different inducer stimulation (29, 30). Monolayer cultures of P19 cells challenged with 50 nM retinoic acid (RA) differentiate to endodermal-like cells (31).

In the current study we show that JNK activity is increased during RA-induced differentiation of P19 cells and the progression of P19 cell differentiation by G₁₂ and G₁₃ is mediated by JNK, but not ERK or p38 MAPK. Expression of the dominant negative form of JNK1 is shown to block RA-induced differentiation of embryonal cells to endoderm, establishing an obligate role of JNK in a model of early mouse embryogenesis.

EXPERIMENTAL PROCEDURES

The P19 cells were obtained from the American Type Culture Collection (Rockville, MD). Cells were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (HyClone, Logan, UT) in 6% CO₂ humidified chamber. To induce endodermal differentiation, P19 cells were cultured as monolayers on tissue culture plates in growth media supplemented with 10 nM RA (all-trans; Sigma) for 2-5 days.

The antisense sequences, 5-AGCTTCAGGCCGCGGCCGCGCCCCGCTGGGGCCCGCGCG CCAT-3 and 5-AGCTTTGCCGCCGCCGCCGCCTCGGCGGGCCCCTGCGGCTCCTAT-3 derived from the cDNA sequences of mouse G₁₂ and G₁₃ (32), respectively, were engineered into the HindIII/ClaI sites of the pLNCX retroviral vector using standard recombinant DNA techniques (33). Plasmids (34) harboring the constitutively activated form of G₁₂ (pcDNA3-G₁₂Q229L) or of G₁₃ (pcDNA3-G₁₃Q226L) were obtained from Dr. Gary L. Johnson (National Jewish Center for Immunology, Denver, CO). The plasmid containing the cDNA for the Flag epitope-tagged, dominant negative form of JNK1 replaces the dual phosphorylation sites Thr-Pro-Tyr with Ala-Pro-Phe (pCMV5-JNK1APF) and was employed to probe the role of JNK in P19 cell differentiation. P19 cells were transfected with plasmids using Lipofectin^TM (Life Technologies, Inc.), according to the manufacturer's protocol. For pCMV5-JNK1-(APF), the pCW1-neo plasmid was co-transfected to provide a selectable marker. Positive transfectants were selected using G418 (400 mg/ml, Life Technologies, Inc.) and stable transfectant clones maintained in media supplemented with 100 mg/ml G418. The ability of the antisense RNAs to suppress the expression of G₁₂ and G₁₃ in the transfectant clones was determined by immunoblotting analysis. Expression of pcDNA3-G₁₂ (Q229L) or pcDNA3-G₁₃ (Q226L) was established by reverse transcriptase-polymerase chain reaction.

For immunoprecipitations and FPLC analysis, P19 cells were washed with phosphate-buffered saline (pH 7.2) and lysed. Lysis buffer contains 150 mM NaCl, 5 mM EDTA, 50 mM NaF, 40 mM sodium pyrophosphate, 50 mM KH₂PO₄, 10 mM sodium molybdate, 20 mM Tris-HCl (pH 7.4), 5 mg/ml aprotinin, 5 mg/ml leupeptin, 6.0 mM dithiothreitol, 2 mM sodium orthovanadate, 200 mM phenylmethylsulfonyl fluoride, 0.1% sodium dodecyl sulfate, and 1% Triton X-100. After 20 min at 4 °C with constant rotation the cell lysate was subjected to centrifugation at 14,000 × g for 15 min, and the resultant supernatant was transferred to a fresh tube. Protein content was measured by the Lowry method (35), using bovine serum albumin as the standard.

Aliquots of protein were subjected to 10% SDS-polyacrylamide gel electrophoresis, and the separated proteins were transferred electrophoretically from the gel to nitrocellulose membrane. Antibodies to the following antigens employed in these studies were obtained from the indicated sources: JNK 1, Santa Cruz Biotechnology, Santa Cruz, CA; ERK1 and 2, Zymed Laboratories Inc., San Francisco; p38 MAPK, Upstate Biotechnology, Lake Placid, NY, G₁₂, Dr. Gunther Schultz, Freie University, Berlin, Germany and Santa Cruz Biotechnology, Santa Cruz, CA; and G₁₃, Dr. Gunther Schultz, Freie University, Berlin, Germany and Santa Cruz Biotechnology, Santa Cruz, CA, or a rabbit polyclonal antiserum obtained by immunization with a synthetic C-terminal dodecapeptide of G₁₃. The immune complexes formed were made visible by using alkaline phosphatase-linked goat anti-rabbit IgG (Life Technologies, Inc.), rabbit anti-mouse IgG (Life Technologies, Inc.), or rabbit anti-goat IgG (Sigma).

An aliquot (250 µg to 1 mg of protein) of the clarified lysate was mixed with the JNK1-specific antibody and resuspended to 1.0 ml in the same cell lysis buffer. After 2 h at 4 °C with constant rotation, 20-40 µl of Protein A/G Plus-agarose (Santa Cruz Biotechnology, Santa Cruz, CA) was added, and the incubation was continued for an additional 1.5 h. The immunoprecipitates were pelleted by centrifugation at 14,000 × g for 5 min. The pellets were washed twice with same lysis buffer and then washed twice with JNK assay buffer (20 mM HEPES (pH 7.5), 20 mM MgCl₂, 0.1 mM Na₃VO₄, 2 mM dithiothreitol, 20 mM -glycerophosphate). Immunoprecipitates were resuspended in a final volume of 40 µl of the JNK assay buffer. A 20-µl aliquot was used for Western blot analysis and the other 20 µl was used for JNK assay. Aliquots of immunoprecipitates were mixed with 2 µg of GST-fusion proteins encoding for the N-terminal region of Jun-(1-79) (36) and placed on ice. Reaction was started by addition of [-³²P]ATP (10 µCi/tube, 100 µM final concentration) and incubated for 15 min at 30 °C. Reaction was stopped by addition of 6 µl of 4 × Laemmli buffer (200 mM Tris-HCl (pH 6.8), 0.4 mM dithiothreitol, 8% SDS, 0.08% bromphenol blue, and 40% glycerol). After separation of proteins by 10% SDS-PAGE and staining with Coomassie Brilliant Blue R250 and autoradiography, the bands for GST-Jun were excised and the incorporated ³²P was quantified by liquid scintillation counting.

Undifferentiated and differentiated P19 cells were harvested and lysed. An aliquot of sample (2.5 mg of protein) was resuspended in 1 ml of lysis buffer and fractionated by chromatography on a Mono Q HR 5/5 FPLC column (Pharmacia Biotech Inc., Uppsala, Sweden). The column was washed with 8 column volumes of buffer A (70 mM -glycerophosphate (pH 7.2), 0.1 mM Na₃VO₄, 2 mM MgCl₂, 1 mM EGTA, and 1 mM dithiothreitol), and the bound proteins were eluted with a 12.5-ml linear gradient of NaCl (0-400 mM) in buffer A. Twenty-five fractions (0.5 ml) were collected at a flow rate of 1 ml/min.

A 15-µl aliquot of FPLC fraction was mixed with 5 µl of 4 × kinase assay buffer (100 mM -glycerophosphate, 160 mM HEPES (pH 7.2), 0.2 mM Na₃VO₄, 40 mM MgCl₂, 100 µg/ml protein kinase A inhibitor peptide, 2 mM EGTA) and 2 µg of either MBP (Sigma) or GST-ATF2-(1-109) (36) used as substrates. Each reaction was started by addition of [-³²P]ATP (10 µCi/tube, 100 µM final concentration) and incubated for 15 min at 30 °C. The reactions were terminated by addition of 6 µl of 4 × Laemmli buffer. After separation of proteins by SDS-PAGE on 12% acrylamide gels, staining of the resolved proteins with Coomassie Brilliant Blue R250, and autoradiography, the bands for either MBP or GST-ATF2-(1-109) were excised, and the incorporated ³²P into these substrates was quantified by liquid scintillation counting.

Cells were grown in Lab-Tek chamber slides (Nunc, Rochester, NY) and were fixed for 10 min with 3% paraformaldehyde. After fixation, slides were rinsed three times with modified Shield's media/Pipes (modified Shield's media; 18 mM MgSO₄, 5 mM CaCl₂, 40 mM KCl, 24 mM NaCl, 5 mM Pipes (pH 6.8), 0.5% Triton X-100, 0.5% Nonidet P-40). A monoclonal antibody MC-480 (SSEA-1) that reacts with a stem-cell surface antigen marker (37) was used to identify the embryonal-stage cells. MC-480 was obtained from the NIH Developmental Studies Hybridoma Bank (Baltimore, MD). The fixed cells were incubated with primary antibody for 30 min at 37 °C and then washed three times with modified Shield's media/Pipes buffer. Fluorescein-conjugated goat anti-mouse (Kirkegaard & Perry Laboratories, Gaithersburg, MD) was added and incubated for 30 min at 37 °C. The slides were washed then three times with blotting buffer (560 mM NaCl, 10 mM KPO₄ (pH 7.5), 0.1% Triton X-100, 0.02% SDS). The cells were examined by phase contrast and epifluorescence microscopy on a Zeiss Axiphot microscope and photographed on Kodak-64T or Kodak-TMax 400 film.

RESULTS AND DISCUSSION

The P19 cells differentiate to neuroectoderm-like phenotypes (29, 30) when allowed to aggregate in the presence of 100 nM retinoic acid. In the absence of aggregation, P19 cells are induced to differentiate by RA but only to the endoderm-like phenotype (29, Fig. 1), providing a useful model for study of this process. P19 embryonic cells stain positive in indirect immunofluorescence study of cells probed with antibody to the embryonic marker SSEA-1 (Fig. 1, panel B). The cells induced morphologically to endoderm by RA no longer stain positive with the same antibody, displaying a loss in the embryonic marker antigen SSEA-1 (Fig. 1, panel D). Endodermal cells stain positive with TROMA-1, a monoclonal antibody to the endoderm-specific marker cytokeratin-1 (37, 38). Untreated P19 cells stain negative for TROMA-1 (Fig. 1, panel F), reflecting their embryonal character, whereas RA-treated P19 cells stain prominently for TROMA-1 (Fig. 1, panel H), a hallmark for endodermal development (37, 38). The absence of epifluorescence signals from the RA-treated cells stained with SSEA-1, from the untreated P19 cells stained with TROMA-1, and from the controls performed without primary antibody (not shown) provide ample evidence for the specificity of the staining in defining endoderm versus embryonal character (Fig. 1). Since heterotrimeric G-proteins have been implicated in control of P19 cell differentiation (39) and G₁₂ and G₁₃ are known to mediate the activation of mitogen-activated protein kinases, especially c-Jun amino-terminal kinase (16, 40), we explored the role of MAP kinases in P19 cell differentiation, with an emphasis on G₁₂ and G₁₃.

To explore the role of MAP kinases in P19 cell differentiation, cells were treated with retinoic acid (100 nM) for 3 days to achieve full differentiation to endodermal phenotype and a cell extract prepared and fractionated by FPLC on a Source Q column (Fig. 2A). Immunoblotting of samples from the fractions revealed ERK1,2 in fractions 9-11 and p38 in fractions 14-19 of extracts from untreated P19 cells. MAP kinase activity was measured using myelin basic protein (MBP), which is a substrate for both p38 and ERK1,2. The profile of activities coincides with the appearance of ERK1,2 and p38 kinases in the immunoblots, shown as insets (Fig. 2, panel A). When MAP kinase activity was measured in fractions from an FPLC of RA-treated, endodermal cells, a modest but highly reproducible (n = 3) decline in ERK1,2 activity was observed in each profile (Fig. 2, panel A). MAP kinase activity in the region of the profile in which p38 was found increased slightly in the extracts from the RA-treated, endodermal P19 cells in each of three separate analyses (Fig. 2, panel A). The status of p38 kinase activity in the FPLC fractions from untreated and RA-treated P19 cells was measured further, using GST-ATF2 fusion protein, as ATF2 is a specific substrate for p38 and not ERK1,2 (Fig. 2B). The activity was confined to the same FPLC fractions as judged by immunoblotting and phosphorylation of the ATF2-fusion protein. In extracts from RA-treated P19 cells, again, the p38 activity displayed a small but reproducible increase compared with the activity in profiles from untreated P19 cells.

Unlike the modest changes observed in the activity of p38 or ERK1,2 in the RA-treated cells, c-Jun amino-terminal kinase activity (JNK) displays marked activation upon treatment of the cells with RA (Fig. 2C). JNK activity in the solid-state assay using GST-Jun fusion protein increases from 2.7-fold on day 2 to 7.8-fold on day 3 following induction of differentiation in response to RA. Based upon multiple experiments, the increase in JNK activity (fold over basal) was found to be 4.3 ± 0.6 (n = 3) on day 2 and 7.0 ± 0.2 (n = 4) on day 3 following treatment with RA. The amount of the JNK itself, in contrast, was found to be relatively constant over the time frame of the induction of differentiation, whereas the activity of JNK was markedly enhanced. These data demonstrate among the various MAP kinases that JNK is markedly activated in the P19 cells induced to endoderm by RA, p38 activities increase little, and ERK1,2 activities show a modest decline.

Anisomycin is a potent activator of JNK activity (15) and was employed to explore the nature of the activation of JNK that accompanies RA treatment (Fig. 3). P19 cells were incubated with or without anisomycin (100 ng/ml) for 2 h, and the JNK was immunoprecipitated from the cell extracts and the kinase activity measured using the solid-state assay (Fig. 3A). Anisomycin proved to be a potent stimulator of JNK, increasing JNK activity by more than 3-fold within 2 h of stimulation. The stimulation by anisomycin was reversible, simply removing the anisomycin precipitated a return to base line of the JNK activity in cells treated with anisomycin, washed free of the agent, and measured 1 day later (Fig. 3A). Anisomycin treatment alone was not able to induce the loss of SSEA-1 antigen nor positive staining by TROMA. The same situation was not observed with regard to the ability of RA to induce differentiation and activation of JNK in the P19 cells (Fig. 3B). Treating cells with 100 nM RA activates JNK by at least 6-fold by day 5, whether the RA is included throughout the 5-day period or added only for the first 2 of the 5-day treatment. The RA-induced increase in JNK activity is sustained following disruption of P19 cell aggregates, harvest, and re-plating of the cells (data not shown). These data suggest a fundamental difference in the activation of JNK by RA as compared with anisomycin. Commitment to the differentiated state or activation of JNK observed in response to RA in the first 2 days continues well beyond the presence of the RA, much like the expression of G₁₂ and G₁₃ induced by RA in P19 cells (39).

RA treatment increases the expression of G₁₂ and G₁₃ in P19 cells (39). Both G₁₂ and G₁₃ have been shown to mediate signals that act on downstream MAP kinases, particularly JNK (16, 40, 41). Taken together these observations stimulated us to examine the relationship between G₁₂ and G₁₃ levels in P19 cells and the activity of JNK (Fig. 4). P19 cells stably expressing mutant forms of either G₁₂ (Q226L) or G₁₃ (Q229L) that are constitutively active were probed with respect to JNK activation, in the absence of RA treatment. Expression of either Q226LG₁₂ or Q229LG₁₃ in P19 cells resulted in the activation of JNK activity in the absence of RA treatment. Cells expressing Q226LG₁₂ displayed activation of JNK activity ranging from 1.9- to 2.7-fold over basal levels. P19 cells expressing Q229LG₁₃ displayed activation of JNK activity ranging from 2.4- to 3.8-fold over basal (Fig. 4). In the absence of RA treatment, P19 cells expressing Q229LG₁₃ have been shown to differentiate to the endodermal phenotype (39). Expression of Q226LG₁₂ induces the differentiation from embryonal phenotype and loss of positive staining for the SSEA-1 embryonic-specific marker but not to endodermal-like phenotype (39). These data are consistent with a role for G₁₂ and G₁₃ controlling aspects of differentiation induced by morphogens (38) but go further to demonstrate a potential role for activation of the MAP kinase regulatory network, and specifically JNK, by RA.

Expression of either constitutively active G₁₂ (Q229L) or G₁₃ (Q226L) activates c-Jun amino-terminal kinase in P19 embryonal carcinoma cells.

The linkage between expression of G₁₂ and G₁₃ and JNK activity in differentiating P19 cells was approached from the opposite perspective, i.e. suppression of G₁₂ and G₁₃ with RNA antisense to each. Previously we demonstrated the suppression of either G₁₂ or G₁₃ in stable transfectants of P19 cells using the pLNCX vector (39). Suppression of G₁₂ was 50-70% in clones expressing RNA antisense to G₁₂ (39), whereas induction of JNK activity upon RA treatment in those clones was attenuated significantly in comparison to clones harboring the empty vector (Fig. 5) or to wild-type cells (Fig. 2C). The same basic observations were true for the clones in which RNA antisense to G₁₃ suppressed G₁₃ levels 60-80%, and induction of JNK activity upon RA treatment was found to be reduced by approximately half. Thus, expression of constitutively active forms of G₁₂ and G₁₃ results in constitutively elevated basal levels of JNK activity, whereas suppression of G₁₂ and G₁₃ results in constitutively attenuated induction of JNK activity upon RA treatment.

Suppression of either G₁₂ or G₁₃ by antisense RNA attenuates the activation of c-Jun amino-terminal kinase in P19 embryonal carcinoma cells.

To test further the linkages among G₁₂ and G₁₃, JNK activation and induction of differentiation by RA, we stably expressed a dominant negative mutant form of JNK1 (JNK1 (APF)) in P19 cells and examined JNK activity of the clones in the absence and presence of RA (Fig. 6A). Densitometric scanning of immunoblots reveals that the total level of JNK1 immunoreactivity increased about 50% in the cells stably expressing the immunoreactive, dominant negative form of JNK1 (not shown). The FLAG-tagged, dominant negative form of JNK1 was readily detected in the cells transfected with the vector harboring JNK1 (APF) with antibodies specific to the tag epitope (not shown). More importantly, expression of negative dominant JNK1 was observed to block completely the activation of JNK1 activity by RA, although slightly increasing the apparent level of JNK basal activity (Fig. 6A). Thus, expression of dominant negative JNK abolishes the ability of RA to activate JNK.

To test the ability of the dominant negative to block endogenous JNK1 activity, we probed activity of JNK following anisomycin activation in cells stably expressing either pCMV vector alone or pCMV vector harboring JNK1 (APF). Upon treatment with 200 ng/ml anisomycin, control cells displayed a 15-fold activation of JNK over basal levels, whereas cells expressing the dominant negative showed a 4-fold activation (Fig. 7B). These data strongly suggest that induction of JNK activity is a causal effect for differentiation rather than result of differentiation.

The capacity to abolish the activation of JNK in response to RA by expressing the dominant negative form of JNK1 provided the opportunity to address directly if the activation of JNK is critical to the ability of RA to induce differentiation of the embryonal P19 cells. Clones harboring the empty vector (pCMV) as well as those expressing the dominant negative version of JNK1 were both treated with RA for 3 days and then stained with antibody to the embryonic-specific marker SSEA-1 (Fig. 7). When treated with RA, clones harboring the empty vector differentiated to the endodermal phenotype just as wild-type P19 cells (Fig. 1), displaying negative staining for the embryonal marker antigen, SSEA-1 (Fig. 7). Clones expressing the dominant negative from of JNK1, in contrast, stain positive for the embryonal marker in the absence or presence of RA. These data demonstrate an obligate role for JNK1 activation in the differentiation of P19 cells to endoderm stimulated by RA.

P19 embryonal carcinoma cells are an attractive model for the study of differentiation, displaying a capacity to differentiate to endoderm, neuroectoderm, and beating cardiac myocytes under the appropriate culture conditions (29, 30). The P19 cell line was adopted for the study of the role of heterotrimeric G-protein subunits G₁₂ and G₁₃ as well as MAP kinases in the differentiation of these embryonal cells to endoderm in response to RA. G₁₂ and G₁₃ are expressed in P19 cells, and their levels of expression are increased in response to RA (39). Moreover, expression of the constitutively active mutants of G₁₂ and G₁₃ provoke differentiation from the embryonal phenotype in the absence of RA (39). The role of heterotrimeric G-proteins in the regulation of the MAP kinase regulatory network has been appreciated recently, and the current study illuminates the role of G₁₂ and G₁₃ in differentiation mediated via specific activation of JNK but neither ERK1,2 nor p38. The JNK pathway is required for embryonic viability (42). Transgenic mice in which the gene for MKK4, the upstream activator of JNK, has been interrupted die prior to day 11 of embryonic development (42). This lack of viability is consistent with a role of JNK in mammalian development, perhaps differentiation (42). In Drosophila, the JNK pathway is required for embryonic development (43-45). Embryonic death occurs because of a failure of the migration of the dorsal epithelial cells required for dorsal closure. The early nature of this defect precludes analysis of the role of JNK in differentiation in this system (43-45). The most compelling evidence for the role of JNK1 in P19 embryonal cell differentiation was the ability of the dominant negative mutant of JNK1 to block endodermal differentiation in response to RA. Perhaps these heterotrimeric G-proteins activate JNK via known downstream elements (46) or perhaps through a tyrosine kinase that is regulated directly by G-proteins, as recently reported in an avian lymphocyte cell line (47).