Gα12 and Gα13 Mediate Differentiation of P19 Mouse Embryonal Carcinoma Cells in Response to Retinoic Acid*

P19 mouse embryonal carcinoma cells can be stimulated to differentiate into endodermal-like, mesodermal-like, and neuronal-like phenotypes in response to specific morphogens. At low concentrations, retinoic acid stimulates P19 embryonal cells to differentiate to cells displaying an endodermal phenotype, whereas at higher concentrations it stimulates differentiation to neuroectoderm. The Gα12 and Gα13 subunits of heterotrimeric G-proteins are expressed in the embryonal P19 cells and stimulated in response to retinoic acid as the cells differentiate to endodermal or neuroectodermal phenotypes. Suppression of the expression of either Gα12 or Gα13 by antisense RNA is shown to promote cell detachment from substratum and apoptosis. Expression of the constitutively active, mutant form of Gα12 (Q229L), in contrast, stimulates loss of the embryonal phenotype. Expression of the constitutively active form of Gα13 (Q226L) stimulates differentiation of the cells from embryonal to endodermal, in the absence of retinoic acid. Thus, both Gα12 and Gα13 are essential to stimulation of cell differentiation by retinoic acid. Deficiency of either Gα12 or Gα13 increases programmed cell death.

Heterotrimeric G-proteins 1 play important roles in cell signaling and more complex biological responses such as oncogenesis, cell differentiation, and development (1)(2)(3)(4). Based upon homology of amino acid sequence the ␣-subunits of G-proteins can be segregated into four major families, designated by a prominent member(s) of the group, G␣ s , G␣ i , G␣ q and G␣ 12/13 (5,6). The G-protein subunits G␣ 12 and G␣ 13 , identified most recently, are expressed ubiquitously and encode polypeptides with M r ϳ42,000 (7). Drosophila homolog (concertina) of mammalian G␣ 12 family has been shown to function in a signaling role during gastrulation (8). A similar role for the G␣ 12 family has been proposed in early mouse development, during the production of germ layers in gastrulation (8). Recent work on G␣ 12 and G␣ 13 has focused on the oncogenic potential of G␣ 12 and G␣ 13 , when overexpressed in cells (9 -14). Although it has been shown that disruption of the gene encoding G␣ 13 results in intrauterine death of transgenic mice (15), specific details of the roles of either G␣ 12 or G␣ 13 in cellular differentiation or normal development in mammalian systems are lacking.
P19 embryonal carcinoma cells provide an attractive model for study of early embryonic determination and differentiation (16 -18). Under appropriate culture conditions, P19 cells display the ability to differentiate into derivatives of three germ layers, endoderm, mesoderm, and ectoderm. Treatment of aggregated P19 cells with low concentrations (10 nM) of retinoic acid (RA) leads to differentiation of embryonal cells to endodermal and mesodermal derivatives (17,18). At higher concentrations (100 nM) of RA, P19 cells can be stimulated to form neurons, glial and fibroblast-like cells (17,18).
In the current study we take advantage of the P19 cell system to explore the expression of G␣ 12 and G␣ 13 during RA-stimulated differentiation. Addition of RA increases expression of both G␣ 12 and G␣ 13 , prompting us to examine their role in cellular differentiation via blocking of expression with RNA antisense to G␣ 12 /G␣ 13 and via expression of GTPasedeficient, constitutively active mutant forms of either G␣ 12 or G␣ 13 . G␣ 12 and G␣ 13 are shown to play essential roles in differentiation of P19 cells in response to RA. In the accompanying paper (19), the linkage between G␣ 12 /G␣ 13 and signaling via mitogen-activated protein kinases in differentiation of P19 cells is elucidated.

EXPERIMENTAL PROCEDURES
Cell Culture and Differentiation-The P19 clone was 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% CO 2 humidified chamber. To promote cellular differentiation, cells were cultured either as monolayers on tissue culture plates in normal growth media supplemented with 1 ϫ 10 Ϫ7 M RA (all-trans; Sigma) for 2 to 5 days or as aggregates in bacterial Petri dishes in media supplemented with 10 -500 nM RA. Cells (7 ϫ 10 4 /ml culture media) were plated into bacterial grade Petri dishes and incubated with RA for 2 days. After 2 days, floating aggregates were collected and washed with Dulbecco's modified Eagle's medium and re-plated onto bacterial grade Petri dishes and incubated with RA for an additional 2 days. The aggregates were harvested, dissociated by treatment with phosphate-buffered saline (PBS) containing 1 mM EDTA, and then transferred to tissue culture plates and cultured in medium lacking RA. For F9 embryonic teratocarcinoma cells (F9 cells), cells were cultured on tissue culture plates in Dulbecco's modified Eagle's media supplemented with 15% fetal bovine serum. To stimulate differentiation of these cells to primitive endoderm, the cultures were challenged with 10 nM RA for 4 to 6 days.
Indirect Immunofluorescence Microscopy and Antibodies-For cultures to be studied by indirect immunofluorescence, cells were plated onto Lab-Tek chamber slides (Nunc, Rochester, NY) and when appropriate fixed with 3% paraformaldehyde for 10 min. After fixation, the slides were rinsed three times with modified Shield's media (MSM) buffered with Pipes consisting of 18 mM MgSO 4 , 5 mM CaCl 2 , 40 mM KCl, 24 mM NaCl, 5 mM Pipes (pH 6.8), 0.5% Triton X-100, and 0.5% Nonidet P-40. The following antibodies were used as markers for undifferentiated or differentiated cells: monoclonal antibody MC-480 (SSEA-1), which reacts with a stem-cell surface antigen (20); TROMA-1 an antibody which reacts with the mouse endoderm marker keratin (21); and antibody 2H3, which reacts with the neuroectodermal-specific marker M r ϳ165,000 neurofilament (22). All primary antibodies were obtained from the NIH Developmental Studies Hybridoma Bank (Baltimore, MD). The fixed cells were incubated with primary antibodies for 30 min at 37°C and then washed three times with MSM-Pipes buffer. Fluorescein-conjugated goat anti-mouse IgG (Kirkegaard & Perry Laboratories, Gaithersburg, MD) was added. The mixture was incubated for 30 min at 37°C and then washed three times with blotting buffer (560 mM NaCl, 10 mM KPO 4 (pH 7.5), 0.1% Triton X-100, 0.02% SDS). The fixed cells were examined using an Zeiss Axiophot microscope for both phase contrast and epifluorescence images. The images were photographed on either Kodak-64T or Kodak-TMax 400 film.
RNA Isolation and Northern Blot Analysis-Total cellular RNA was isolated using RNA-STAT 60 reagent (TEL-TEST B, Inc., Friendswood, TX) from undifferentiated and differentiated P19 and F9 cells. A 20-g aliquot of total RNA was separated on 1.2% formaldehyde-containing agarose gel, transferred by electroblotting to Nytran membrane, crosslinked by UV radiation (Stratalinker, Stratagene, La Jolla), and then probed by hybridization with 32 P-labeled cDNA of one of the following: mouse G␣ 12 , mouse G␣ 13 , or rat GAPDH. The DNA probes were labeled by the random-priming method (23), in the presence of [␣-32 P]dCTP.
Plasmids and Stable Transfection-The antisense sequences, 5Ј-AGCTTCAGGCCGCGGCCGCGCCCCGCTGGGGCCCGCGCGCCA-T-3Ј, and 5Ј-AGCTTTGCCGCCGCCGCCGCCTCGGCGGGCCCCTGC-GGCTCCTAT-3Ј derived from the cDNA sequences of mouse G␣ 12 and G␣ 13 (5), respectively, were engineered into the HindIII/ClaI sites of the pLNCX retroviral vector using standard recombinant DNA techniques (24). Plasmids harboring constitutively active mutant forms (25) of either G␣ 12 (pcDNA3-G␣ 12 Q229L) or G␣ 13 (pcDNA3-G␣ 13 Q226L) were provided by Dr. Gary L. Johnson (National Jewish Center for Immunology and Respiratory Medicine, Denver, CO). The P19 cells were transfected with these plasmids using Lipofectin TM (Life Technologies, Inc.), following the manufacturer's protocol. Positive transfectants harboring the neo resistance gene were selected and cloned in the presence of G418 (400 mg/ml, Life Technologies, Inc.). Stable transfectants were maintained in media containing 100 mg/ml G418. The ability of the antisense RNA to suppress the expression of G␣ 12 and G␣ 13 in the stable transfectants was determined by immunoblotting with antibodies specific for each subunit. The expression of the constitutively active mutant forms of the G␣ 12 (Q229L) and G␣ 12 (Q229L) was analyzed by RT-PCR.
Preparation of Cell Membrane Fractions and Total Cell Lysates-Crude cell membrane fractions from the different clones were prepared FIG. 1. P19 embryonal carcinoma cells differentiate to endodermal phenotype in response to retinoic acid. Differentiation was stimulated by allowing the cells to aggregate in bacterial grade Petri dishes for 4 days in medium supplemented with 10 nM RA for endodermal differentiation and 100 nM RA for neuroectodermal differentiation. After 4 days, the aggregates were dissociated by treatment of PBS, 1 mM EDTA and spread on tissue culture plates for 3 days. Undifferentiated cells were aggregated in medium lacking RA. Cells were fixed using 3% paraformaldehyde, and the cells were subjected to phase contrast and indirect immunofluorescence microscopy, as described under "Experimental Procedures." Indirect immunofluorescence was performed with antibodies specific for either the embryonal marker protein SSEA-1 (panels B and in HME buffer (20 mM HEPES (pH 7.4), 2 mM MgCl 2 , 1 mM EDTA) containing the following protease inhibitors: aprotinin (5 mg/ml), leupeptin (5 mg/ml), and freshly prepared phenylmethylsulfonyl fluoride (200 mM). Cells were disrupted by 20 strokes of a Dounce homogenizer. The homogenate was centrifuged at 500 ϫ g for 10 min to remove nuclei, and the postnuclear supernatant was collected as a whole-cell extract. The supernatant was subjected to centrifugation at 50,000 ϫ g for 30 min. The crude membrane pellets were resuspended and their protein content measured by the Lowry method (26), using bovine serum albumin as the standard.
Immunoblotting Analysis-Aliquots of crude membrane proteins or whole cell extracts were subjected to electrophoresis on 10% SDSpolyacrylamide gels, and the separated proteins were transferred electrophoretically from the gel to nitrocellulose membrane. Antibodies specific for G␣ 12 were kindly provided by Dr. Gunther Schultz (Freie University Berlin, Germany) or purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies specific for G␣ 13 were provided as a gift from Dr. Gunther Schultz, purchased from Santa Cruz Biotechnology, or were prepared by the laboratory using a synthetic, dodecapeptide C-terminal peptide as an immunogen in New Zealand White rabbits. The immunecomplexes formed were made visible by secondary staining either with alkaline phosphatase-linked goat anti-rabbit IgG (Life Technologies, Inc.) or with the chemiluminescence system (RENAISSANCE TM , NEN Life Science Products).
Terminal Transferase Assay-To detect apoptotic cells, we employed the terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling technique (27). Cells grown on gelatin-coated coverslips were cytocentrifuged (150 ϫ g for 5 min) and fixed with 3% paraformaldehyde in MSM-Pipes buffer for 15 min at room temperature. The fixed cells were washed with MSM-Pipes buffer, transferred in ice-cold 70% ethanol, and maintained at Ϫ20°C (up to 3 days). After washing with PBS, cells were incubated in 100 l of solution containing 0.1 M potassium cacodylate (pH 7.2), 2 mM CoCl 2 , 0.2 mM dithiothreitol, 20 mM biotin-dCTP (Life Technologies, Inc.), and 30 units of terminal deoxynucleotidyltransferase (Life Technologies, Inc.) for 2 h at 37°C. The slides were then rinsed in PBS and incubated in the dark for 2 h with the staining buffer composed of 10 mg/ml fluoresceinated avidin (Boehringer Mannheim), 4 ϫ SSC (0.6 M NaCl, 0.06 M sodium citrate), 0.1% Triton X-100, and 5% non-fat dry milk. The slides were then washed three times with blotting buffer and examined using a Zeiss Axiophot microscope for phase contrast and epifluorescence. The images were photographed on Kodak-64T film.

RESULTS AND DISCUSSION
P19 mouse embryonal carcinoma cells are sensitive to aggregation and RA, differentiating to an endoderm phenotype in response to low concentrations of RA (10 nM) and to a neuroectodermal-like phenotype in response to high concentration of RA (100 nM). Aggregation on bacterial-grade Petri dishes for 4 days followed by dissociation and 3 days further growth on tissue culture plates is required for stimulation of the differentiated states by RA (18). Untreated P19 cells grow densely packed, display a characteristic cuboidal morphology (panel A) evident in phase contrast microscopy, and are stained prominently by the embryonic marker SSEA-1 (panel B), as shown by indirect immunofluorescence signal (Fig. 1). Treatment with either 10 or 100 nM RA results in a dramatic change in morphology (compare panels C, E, and G with panel A) as well as a loss of SSEA-1 antigen staining (panel D). These data demonstrate the loss of the SSEA-1 antigen in response to RA. Positive staining with TROMA-1, a monoclonal antibody directed against a class of cytokeratin-like proteins, is a hallmark of endodermal-like phenotype. Untreated, embryonal P19 cells fail to stain with TROMA-1 antibody (not shown), whereas P19 cells differentiated to endoderm by treatment with 10 nM RA display prominent staining with TROMA-1 (panel F). At higher concentrations of RA (100 nM), P19 cells differentiate to a neuroectodermal phenotype with neurite-like morphology as well as prominent staining by antibody 2H3 (panel H), an antibody directed against the 165-kDa neurofilament protein (22). The embryonal cells and those treated with 10 nM RA are negative for staining by the neuronal-specific marker, 2H3 antibody (not shown). Thus, P19 cells provide an invaluable model system in which to study how a morphogen like RA, in combination with aggregation, stimulates commitment of these cells to endodermal and neuroectodermal phenotypes.
The expression of G␣ 12 and G␣ 13 mRNAs in P19 cells was days. These same blots were probed with GAPDH, demonstrating equal loading of RNA samples. Panel C, RT-PCR amplification from total RNA prepared from mouse embryos at 6.5 and 7.5 days p.c. and maternal RNA. To preclude maternal contamination of the embryonic RNA, the ectoplacental cone was removed in 7.5-day p.c. embryos. Maternal RNA was isolated from the decidual tissue. RT-PCR was performed with primers specific for amplification of G␣ 12 and G␣ 13 sequences using the RNA samples indicated as templates. Each of the experiments was replicated at least twice with similar quantitative results. investigated in cells allowed to aggregate for 4 days in either the absence or the presence of increasing concentrations of RA and then dissociated and grown further for 3 days on tissue culture plates (Fig. 2). Northern analysis detected transcripts for both mRNAs (ϳ4 kilobase pairs for G␣ 12 and ϳ6 kilobase pairs for G␣ 13 ) in the wild-type, untreated P19 cells ( Fig. 2A), the G␣ 12 mRNA appearing more abundant than that of G␣ 13 .
Although mRNA levels of glyceraldehyde phosphate dehydrogenase (GAPDH) were relatively constant in P19 cells challenged with RA from 10 to 500 nM concentrations, transcripts for both G␣ 12 and G␣ 13 were found to increase 2.5-2.9-fold over levels in untreated cells upon aggregation in the presence of high concentrations of RA for 4 days, followed by dissociation and re-growth on tissue culture plates. The increase in G␣ 12 and G␣ 13 mRNA levels prompted us to explore if similar challenge of another embryonal-derived cell line, the mouse F9 teratocarcinoma cells, would stimulate expression of either G␣ 12 or G␣ 13 transcripts, or both (Fig. 2, panel B). Like the P19 cells, F9 teratocarcinoma cells displayed a 2.8-fold increase in G␣ 12 and 1.8-fold increase in G␣ 13 mRNA levels after 4 days of challenge with RA. At 6 days of exposure to RA, levels of G␣ 12 mRNA increase by 2.3-fold and those of G␣ 13 mRNA by 2.6-fold over that of untreated, F9 embryonal stem cells in culture. Expression of mRNA levels for GAPDH, in contrast, was relatively constant in response to RA. Under these conditions, RA stimulates F9 cells to differentiate from embryonal stages to primitive endoderm (28). These data suggest that commitment of embryonal cells to the differentiated state in response to RA is accompanied by a sharp rise in the expression of G␣ 12 and G␣ 13 , at the level of mRNA.
We next addressed if transcripts for G␣ 12 and for G␣ 13 could be detected in mouse embryos. Reverse transcriptase-polymerase chain reaction (RT-PCR) amplification with primers specific for G␣ 12 and for G␣ 13 reveals the presence of both G␣ 12 and G␣ 13 transcripts in mouse embryos harvested at days 6.5 and 7.5 post-coitus (Fig. 2, panel C). RT-PCR product was observed also in RNA extracts of the maternal tissue decidua. To avoid contamination of maternal RNA, we manually dissected the ectoplacental cone from the 7.5-day p.c. embryos. The stimulation of G␣ 12 and G␣ 13 gene expression in both P19 and F9 cells as well as the detection of G␣ 12 and G␣ 13 mRNAs in early stage of mouse embryo prompted us to investigate if G␣ 12 and G␣ 13 may play a role in cellular differentiation and/or mouse embryogenesis. G␣ 12 and G␣ 13 expression was explored first at the protein level by immunoblotting (Fig. 3). Crude membranes from un- The membranes (200 mg protein/lane) were separated SDS-polyacrylamide gel electrophoresis on 10% gels, transferred to nitrocellulose membranes, and probed with primary antibodies to either G␣ 12 (upper) or to G␣ 13 (lower). Immunoblots stained for G␣ 12 were made visible by alkaline phosphatase-conjugated secondary antibody. Immunoblots stained for G␣ 13 were made visible by use of the chemiluminescence regent (RENAISSANCE TM , NEN Life Science Products). Data shown are representative of at least three independent experiments performed on separate occasions.

FIG. 4. Expression of RNA antisense to G␣ 12 and G␣ 13 stimulates a decline in subunit expression in stable transfected, P19 embryonal carcinoma cells: analysis by immunoblotting.
Crude cell membranes were prepared from stable transfectant clones expressing antisense RNA specific for either G␣ 12 (lanes 2 and 3, panel A) or G␣ 13 (lanes 2-4, panel B) as well as clones harboring the empty pLNCX vector (lane 1, panels A and B). Crude membranes (50 mg protein/lane) were subjected to SDS-polyacrylamide gel electrophoresis on 10% gels, the separated proteins transferred to nitrocellulose blots, and the blots then probed with primary antibodies to either G␣ 12 or G␣ 13 . Immunoblots stained for G␣ 12 were made visible by alkaline phosphataseconjugated secondary antibody. Immunoblots stained for G␣ 13  treated P19 cells (where TC indicates cells grown on tissue culture plates alone) as well as from cells allowed to aggregate in the presence of 10, 100, and 500 nM RA for 4 days (aggregation ϩ TC) and then dissociated and re-grown on tissue culture plates for 3 days were subjected to SDS-polyacrylamide gel electrophoresis and transferred electrophoretically to nitrocellulose. The blots were stained with antibodies specific to either G␣ 12 or G␣ 13 . In good agreement with the analysis of mRNA levels, expression of both G␣ 12 and G␣ 13 was observed in the untreated P19 cells. Challenging cells with RA increases expression of G␣ 12 to levels about 1.5-fold greater than the cells grown on tissue culture plates alone. For G␣ 13 , expression was increased about 2-fold in the cells challenged with RA at either 10 nM (endodermal) or 100 nM (neuroectodermal). At 500 nM RA, levels of G␣ 13 expression rose to more than 2.5-fold. Thus, RA stimulates differentiation of the embryonal carcinoma cells to endoderm or to neuroectoderm and increases the expression of both G␣ 12 and G␣ 13 . Expression of many G-protein subunits, except G␣ s , remains relatively constant during adipogenesis in 3T3-L1 embryonal fibroblasts (3). Similarly, in F9 teratocarcinoma cells differentiating to primitive or parietal endoderm, the levels of G␣ s , G␣ i1 , G␣ i3 , G␣ o , and G␤ 1 all remain relatively constant. G␣ i2 expression uniquely declines (4), whereas levels of G␣ 12 and G␣ 13 are shown to increase (present study) in the F9 model of early mouse development. Thus, generalized changes in G-protein subunit expression are not typical during differentiation in several well known models. Not only have changes in the expression of specific subunits of G-proteins been shown to be associated with cellular differentiation but have also been demonstrated to be critical features in regulating 3T3-L1 cell adipogenesis (3) and F9 teratocarcinoma cell differentiation (4).
To explore whether or not a linkage exists between the expression of either G␣ 12 or G␣ 13 and the capacity of the P19 cells to differentiate in response to RA, expression of the individual ␣-subunits was suppressed via constitutive expression of RNA antisense to either G␣ 12 or G␣ 13 . The vector employed, a derivative of the pLNCX vector (4), takes advantage of the cytomegalovirus promoter and has been shown to express antisense RNA and to suppress targeted protein expression in F9 teratocarcinoma cells (4), as well as other cell types like hamster vas deferens smooth muscle DDT1 MF-2 and human epidermoid carcinoma cells (29). Although this approach has succeeded in suppressing G i ␣ 2 to near absence in several cells (4), we identified no P19 cell clones in which either G␣ 12 or G␣ 13 expression was abolished (Fig. 4). After careful analysis of many clones, we observed that suppression of G␣ 12 by antisense RNA was no greater than 70%. Similarly, the level of suppression of G␣ 13 observed in viable clones never exceeded 80%. Clones expressing RNA antisense either to G␣ 12 or to G␣ 13 display prominent staining by antibodies to SSEA-1 (not shown), demonstrating an embryonal character even in the absence wild-type levels of either G␣ 12 or G␣ 13 . Although possible that the vector may not operate well in the P19 cells, ample examples of vector expression in a variety of other cells, including F9 teratocarcinoma cells (4), suggest that this not likely. Alternatively, one can speculate that perhaps the absence of either G␣ 12 or G␣ 13 cannot be tolerated and that the only viable clones are those expressing a threshold level of subunit sufficient to support cell proliferation.
Favoring the second possibility, one that could be tested directly, we examined the growth of clones in which either G␣ 12 or G␣ 13 was suppressed at least by 50 -70% (Fig. 5). Cells expressing RNA antisense to G␣ 12 and levels of G␣ 12 deficiency FIG. 6. Deficiency in expression of either G␣ 12 or G␣ 13 stimulates apoptosis in P19 embryonal carcinoma cells. Cells (ϳ1.0 ϫ 10 5 cells) from stable transfectants made deficient by antisense RNA in either G␣ 12 (ASG␣ 12 , panels C and D) or G␣ 13 (ASG␣ 13 , panels E and F) as well as stable transfectant clones harboring empty vector (pLNCX, panels A and B) were seeded onto 6-well plates containing acid-treated and gelatin-coated coverslips. After 2 days of growth, detached cells were re-attached to the coverslip by cytospin. The cells were fixed with 3% paraformaldehyde for 15 min, incubated in the presence of exogenous terminal deoxynucleotidyltransferase and biotin-14-dCTP, and then counterstained with fluoresceinated avidin. The slides were viewed by phase contrast (panels A, C, and E) and epifluorescence (panels B, D, and F) microscopy. Data shown are representative of at least three independent experiments performed on separate clones and separate occasions. of ϳ60 -70% displayed a 2-fold decline in cell growth and an increase in the percentage of cells that detach from the substratum. For G␣ 12 -deficient clones, cell number declined by two-thirds, whereas cell detachment was approximately 35%. Cells constitutively expressing RNA antisense to G␣ 13 , in contrast, display only a slight reduction in cell growth but a prominent increase in cell detachment from the substratum. Approximately half of the cells detach from the culture dishes in the clones deficient in G␣ 13 . Although trypan blue exclusion was evident in the bulk of the cells detached from the plates for clones deficient in either G␣ 12 or G␣ 13 (not shown), the more direct question is whether or not these cells have initiated programmed cell death, apoptosis. To explore this possibility, clones were fixed with paraformaldehyde and then assayed for apoptosis using exogenous terminal deoxynucleotidyltransferase with biotinylated 14-dCTP, counterstained with fluorescein-coupled avidin (Fig. 6). Although displaying some background staining, P19 cells transfected with vector alone did not show prominent staining using this version of the terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling technique, reflecting few cells entered into apoptosis. Stable transfectants of P19 cells made deficient in either G␣ 12 or G␣ 13 , in contrast, showed marked staining of cells using the dUTP-biotin nick end labeling assay, reflecting marked apoptotic activity. These data clearly link P19 cell deficiency in either G␣ 12 or G␣ 13 to initiation of programmed cell death and detachment.
The increase in G␣ 12 and G␣ 13 expression in RA-treated cells, coupled with the increased cell detachment/apoptosis in clones expressing RNA antisense to either G␣ 12 or G␣ 13 , raises the possible linkage between G␣ 12 and G␣ 13 expression and control of differentiation in P19 embryonal cells. To test the linkage further, P19 cells were stably transfected with an expression vector harboring the cDNA for the constitutively active mutant form of each of the two ␣-subunits. Using the pcDNA3 vector, we created clones stably expressing the constitutively active mutant forms of G␣ 12 (Q229L) and G␣ 13 (Q226L) and characterized their properties. RT-PCR and primers from a unique sequence in either G␣ 12 or G␣ 13 in concert with a primer just 5Ј to the poly(A) region of the expression vector pCDNA3 revealed the presence of the mRNAs encoding the mutant versions of G␣ 12 and G␣ 13 could be detected readily in the stable transfectants but not in the transfectant clones harboring the empty vector (Fig. 7). Control studies of the  1 and 2) or pCDNA3 vector harboring the cDNA for the constitutively active mutant forms of either Q229LG␣ 12 (12QL, lanes [3][4][5] or Q226LG␣ 13 (13QL, lanes 6 -8). Simultaneous amplification was performed with primers specific for GAPDH, demonstrating equivalent loading for lanes 1-8. Lower panel, phase contrast (A, C, E, and G) and epifluorescence (B, D, F, and H) images of stable transfectant clones harboring empty vector (pcDNA3, A and B) or pcDNA3 vector harboring the cDNA for the constitutively active mutant forms of either G␣ 12 Q229L (panels C and D) or G␣ 13 Q226L (panels E--H). The cells subjected to indirect immunofluorescence after staining with antibodies specific for an embryonal marker (SSEA-1, panels B, D, and F) or endodermal marker (TROMA-1, panel H). Negative staining was observed for G␣ 12 Q229L clones probed with antibodies to TROMA-1 (not shown) and for both G␣ 12 Q229L clones and G␣ 13 Q226L clones probed with antibodies to 2H3, the neuroectodermal marker (not shown). Data shown are representative of at least three independent experiments performed on separate clones and separate occasions.
RT-PCR employed primers for GAPDH, displaying equivalent amplification from the samples obtained from clones harboring empty vector as well as those harboring the cDNAs for either G␣ 12 (Q229L) or G␣ 13 (Q226L). Although not quantitative, RT-PCR amplification of RNA from clones transfected with the expression vector harboring the mutant forms of either ␣-subunit consistently displayed prominent signals (Fig. 7).
Clones stably transfected with expression vectors harboring the constitutively active forms of G␣ 12 or G␣ 13 display morphological changes that prompted an investigation of their phenotypes. Cells stably transfected with empty pcDNA3 vector alone display the wild-type P19 cell morphology and stained prominently for the embryonic marker SSEA-1 (compare Figs. 1 and 7). The clones stably transfected with either pcDNA3G␣ 12 (Q229L) or pcDNA3G␣ 13 (Q226L), in contrast, have differentiated from the embryonal phenotype as signified by the essential loss of SSEA-1 expression. In addition, these cells assumed a morphology displayed in the phase contrast microscopy that was consistent with a differentiation from the embryonal phenotype. These same clones were probed by staining with markers for both endoderm (TROMA-1) as well as neuroectoderm (2H3). Clones stably transfected with pcDNA3G␣ 12 (Q229L) were negative for staining for TROMA-1 as well as 2H3 (not shown), although also negative for staining of SSEA-1. Clones stably transfected with pcDNA3G␣ 13 (Q226L) also were negative for staining of SSEA-1 antibody but positive for staining with TROMA-1 antibody, the endodermal marker (Fig. 7). Taken together, these data suggest that the expression of the constitutively active mutant form of G␣ 13 (Q226L) promoted differentiation of embryonal carcinoma cells to endoderm in the absence of stimulation by RA (Fig. 1). Staining of these G␣ 13 (Q226L)-expressing clones with 2H3 was negative, as it was for G␣ 12 (Q229L)-expressing cells (not shown).
The data displaying stimulation of G␣ 12 and G␣ 13 expression during P19 cell differentiation (Figs. 2 and 3) are consistent with those for the regulation of Na ϩ /H ϩ exchanger, NHE-1 (30). NHE-1, the only ubiquitously expressed member of the Na ϩ /H ϩ exchange subtype present in most eukaryotic cells, regulates intracellular pH by controlling exchange of intracellular H ϩ for extracellular Na ϩ (31,32). It has been shown that NHE-1 activity is stimulated during P19 cell differentiation (33) and that activation of NHE-1 is associated with induction of c-fos and c-jun mRNAs (34). Expression of exogenous GTPase-deficient mutants of G␣ 12 (Q229L) or G␣ 13 (Q226L) in COS-1 or NIH 3T3 cells stimulates JNK (35) and NHE-1 activities (31,32,36). The findings presented herein show for the first time that expression of endogenous G␣ 12 and G␣ 13 is increased by retinoic acid. In P19 cells, this stimulation of differentiation by retinoic acid may be linked to NHE-1 activity (33) and perhaps to activation of JNK (19).
Reduction in the cellular complement of either G␣ 12 or G␣ 13 provokes programmed cell death and cell detachment (Figs. 5 and 6). It is not clear if detachment of cells is due to apoptosis or vice versa. Stimulation of stress fiber formation and focal adhesion assembly by the overexpression of G␣ 12 (Q229L) or G␣ 13 (Q226L) has been reported (25), and impaired migratory responses to thrombin have been observed in G␣ 13 (Ϫ/Ϫ) embryonic fibroblasts (15). Based upon these recent studies, deficiency in either G␣ 12 or G␣ 13 might be predicted to alter cell attachment, which was in fact observed. In addition, the cells deficient in either G␣ 12 or G␣ 13 clearly display increased programmed cell death.
RA promotes differentiation of P19 embryonal carcinoma cells to various germ layer-like phenotypes, including endoderm and neuroectoderm. The expression of G␣ 12 and G␣ 13 subunits is increased by retinoic acid, suggesting the possibility that these G-protein subunits are important or even essential elements of RA-stimulated differentiation. This fundamental question was addressed in the present work using a variety of approaches. The evidence linking G␣ 12 and G␣ 13 to the ability of retinoic acid to promote cell differentiation is as follows: both G␣ 12 and G␣ 13 display increased expression in cells undergoing differentiation in response to RA; complete suppression of the subunits by antisense RNA was not observed in viable clones; deficiency of either G␣ 12 or G␣ 13 provoked increased apoptosis and cell detachment; expression of constitutively active G␣ 12 (Q229L) stimulated P19 cells to differentiate from the embryonal stage; and expression of constitutively active G␣ 13 (Q226L) stimulated endodermal differentiation in the absence of RA. These data provide a compelling argument supporting a linkage between differentiation of P19 embryonal cells and expression of G␣ 12 and G␣ 13 .