Gα12 Stimulates Apoptosis in Epithelial Cells through JNK1-mediated Bcl-2 Degradation and Up-regulation of IκBα*

Apoptosis is an essential mechanism for the maintenance of somatic tissues, and when dysregulated can lead to numerous pathological conditions. G proteins regulate apoptosis in addition to other cellular functions, but the roles of specific G proteins in apoptosis signaling are not well characterized. Gα12 stimulates protein phosphatase 2A (PP2A), a serine/threonine phosphatase that modulates essential signaling pathways, including apoptosis. Herein, we examined whether Gα12 regulates apoptosis in epithelial cells. Inducible expression of Gα12 or constitutively active (QL)α12 in Madin-Darby canine kidney cells led to increased apoptosis with expression of QLα12, but not Gα12. Inducing QLα12 led to degradation of the anti-apoptotic protein Bcl-2 (via the proteasome pathway), increased JNK activity, and up-regulated IκBα protein levels, a potent stimulator of apoptosis. Furthermore, the QLα12-stimulated activation of JNK was blocked by inhibiting PP2A. To characterize endogenous Gα12 signaling pathways, non-transfected MDCK-II and HEK293 cells were stimulated with thrombin. Thrombin activated endogenous Gα12 (confirmed by GST-tetratricopeptide repeat (TPR) pull-downs) and stimulated apoptosis in both cell types. The mechanisms of thrombin-stimulated apoptosis through endogenous Gα12 were nearly identical to the mechanisms identified in QLα12-MDCK cells and included loss of Bcl-2, JNK activation, and up-regulation of IκBα. Knockdown of the PP2A catalytic subunit in HEK293 cells inhibited thrombin-stimulated apoptosis, prevented JNK activation, and blocked Bcl-2 degradation. In summary, Gα12 has a major role in regulating epithelial cell apoptosis through PP2A and JNK activation leading to loss of Bcl-2 protein expression. Targeting these pathways in vivo may lead to new therapeutic strategies for a variety of disease processes.

Signaling through G proteins 2 is an important mechanism regulating apoptosis, a highly conserved process of pro-grammed cell death fundamental to normal development and somatic maintenance in all multicellular organisms. Changes in apoptosis signaling pathways contribute to major disease processes including cancer, degenerative neurological diseases, such as Parkinsonism and Alzheimer, and kidney failure from numerous etiologies including polycystic kidney disease. There are two major signaling pathways leading to apoptosis. The intrinsic pathway is mediated by the permeabilization of mitochondria and caspase-9 activation and is a general response to cell damage or stress. The extrinsic pathway is activated by stimulation from specific death inducing factors, such as Fas (apoptosis-stimulating fragment) and tumor necrosis factor and is mediated by caspase-8 (1). In the intrinsic pathway, the Bcl family (including pro-and anti-apoptotic proteins) is central to determining whether cells will undergo apoptosis (reviewed in Ref. 2). Bcl-2 is localized in the outer mitochondrial, endoplasmic reticulum, and perinuclear membranes, and under proliferative and anti-apoptotic (pro-survival) conditions, Bcl-2 heterodimerizes with pro-apoptotic protein BAX to form a stable, non-conductive complex (2). With an apoptotic stimulus, Bcl-2 is phosphorylated; this prevents dimerization with Bax and also leads to Bcl-2 degradation, thus permitting BAX to homodimerize and catalyze the formation of the mitochondrial apoptosis-induced channel. This channel allows the efflux of cytochrome c, triggering Apaf-1 and the caspase cascade leading to apoptosis (2,3).
Heterotrimeric G proteins regulate numerous cellular processes including proliferation, differentiation, junctional assembly, and apoptosis. G proteins consist of G␣ and G␤␥ subunits, and form a stable heterotrimeric complex with GDP bound to the ␣ subunit in the resting state. In canonical G protein signaling, ligand binding to a seven-transmembrane domain receptor results in conformational changes in G␣ that lead to dissociation of GDP and separation from G␤␥. GTP subsequently binds to G␣, and signal transduction occurs through G␣ and G␤␥ subunits until the intrinsic G␣ GTPase activity hydrolyzes GTP to GDP. It is now appreciated that G protein signaling is quite complex with the identification of G proteins in subcellular microdomains and in interactions with numerous regulatory and scaffolding proteins. Several studies have implicated signaling through each of the four heterotrimeric G protein families (G␣ s , G␣ i/o , G␣ q/11 , and G␣ 12/13 ) to regulate apoptosis, but the mechanism(s) are not well defined (4 -9).
G␣ 12 and G␣ 13 regulate fundamental cellular processes that include proliferation (10), transformation (11), tight junction assembly (12,13), directed cell migration (14), and regulation of the actin cytoskeleton (15). G␣ 12/13 are potent oncogenes in some cell types (16,17), and may also stimulate apoptosis. In transiently transfected COS cells, G␣ 12 and G␣ 13 induced apoptosis through activation of MEKK1 and Ask1 in a Bcl-2-dependent manner (18), and in a human adenocarcinoma cell line, a mutant G␣ 13 stimulated apoptosis through JNK (19). However, there is very little known about the role of G␣ 12 in regulating apoptosis in non-transfected cells, and few studies have investigated G protein regulation of apoptosis in epithelial cells. Recently, we identified an interaction of G␣ 12 with the regulatory subunit of the serine/threonine phosphatase protein phosphatase 2A (PP2A) and demonstrated that G␣ 12 stimulated PP2A activity in vitro and in MDCK cells (20,21). PP2A is implicated in many of the same cellular processes that have been described for G␣ 12 (reviewed in Ref. 22). PP2A regulates apoptosis by maintaining dephosphorylated Bcl-2 (23) as well as regulating several other members of the Bcl-2 family. Identification of the G␣ 12 -PP2A interaction prompted us to investigate whether these signaling proteins were linked to apoptosis in epithelial cells. Using a well characterized MDCK cell culture model of inducible expression of G␣ 12 , we report that G␣ 12 is a potent activator of apoptosis in epithelial cells and identify critical regulation of Bcl-2 protein levels by G␣ 12 , PP2A, and JNK. Furthermore, we demonstrate for the first time that activation of endogenous G␣ 12 -coupled signaling pathways stimulates apoptosis in epithelial cells and recapitulates the signaling pathways identified in the inducible G␣ 12 -MDCK cell model.
Flow Cytometry-G␣ 12 -or QL␣ 12 -expressing MDCK cells were grown to confluence and cultured Ϯ dox for 72 h. Adherent cells were collected by trypsinization and pooled with floating cells (1-2 ϫ 10 6 ) by low speed centrifugation, washed with phosphate-buffered saline (PBS), and fixed with 70% ethanol in PBS at 4°C for 30 min. Subsequently, cells were incubated with 1 g/ml RNase A in PBS for 30 min at room temperature, after which cells were stained in 50 g/ml propidium iodide (Invitrogen) in PBS for 30 min at room temperature and analyzed by flow cytometry in the propidium iodide/PE Texas Red channel. 10,000 cells were analyzed per experiment.
DNA Fragmentation Assay-G␣ 12 -or QL␣ 12 -expressing MDCK cells were cultured Ϯ dox for 72 h. Adherent cells were collected by trypsinization and pooled with floating cells by low speed centrifugation, washed with PBS, and lysed with genomic DNA extraction buffer (100 mM NaCl, 10 mM Tris-HCl, pH 8.0, 25 mM EDTA, 0.5% SDS, and 0.1 mg/ml proteinase K). Extracts were incubated for 24 h at 50°C and subsequently for 1 h at 37°C with 1 g/ml RNase A (MP Biomedical, Solon, OH). DNA was extracted with an equal volume of phenol/chloroform (1:1) and precipitated at Ϫ70°C for 24 h with 3 volume equivalents of absolute ethanol. DNA pellets were resuspended in 20 l of 10 mM Tris (pH 7.8), 1 mM EDTA buffer and analyzed by electrophoresis on a 0.5% agarose gel run at 50 V for 4 h. Images were obtained using a Kodak DC290 Zoom Digital Camera (Eastman Kodak).
c-Jun N-terminal Kinase Activity-G␣ 12 -or QL␣ 12 -expressing MDCK cells were cultured Ϯ dox for 72 h. Floating cells and adherent cells were collected by trypsinization for 5 min and low speed centrifugation, washed with PBS, and lysed with JNK Extraction Buffer (Biovision). Lysates were normalized for protein concentration and analyzed according to the manufacturer's instructions (KinaseSTAR JNK Activity Screening Kit, Biovision). Briefly, lysates were incubated with GST-cJun on glutathione-Sepharose beads to pull down total JNK, which was resuspended with 200 M ATP and incubated at 30°C for 30 min. Samples were eluted in Laemmli sample buffer and analyzed by SDS-PAGE and Western blot for phospho-cJun.
siRNA Knockdowns-PP2A catalytic subunit was selectively knocked down using SMARTpool pooled siRNAs (Upstate, Lake Placid, NY) that contains four different siRNAs (catalog number D-001206-13-01). A nonspecific control pool (catalog number D-001206-13-01) was used for the negative control and also contained four different siRNAs. HEK293 cells were transfected with siRNA (200 nM) using Lipofectamine (Invitrogen) following the manufacturer's instructions. After overnight incubation, fresh media was added to the cells, and the cells were cultured an additional 24 h. Lysates were prepared and analyzed by Western blot with goat anti-PP2A catalytic subunit antibody at 1:2000. For thrombin stimulation, cells were serum starved overnight prior to treatment with 2 units/ml. JNK phosphorylation and Bcl-2 expression were examined by Western blot at 8 h, and apoptosis was quantified after 24 h.
Quantification and Statistics-Western blots and agarose gels were scanned using an Epson 1640 desktop scanner and band intensity quantified using NIH Image (Wayne Rasband) after subtracting background and determining linear range. Statistics were done in GraphPad Prism (San Diego, CA). Significance was determined by using the two-tailed t test.

Expression of QL␣ 12 in MDCK-II Cells Induces Apoptosis-
Several factors including the disruption of cell-matrix interactions and dome formation associated with confluence have been shown to trigger apoptosis in MDCK cells (25,26). The pathways associated with these phenotypes are not known. The study of G␣ 12 has been limited by low levels of expression in cells, and by its novel protein characteristics that make it difficult to express in vitro and in cell culture models. We have established and extensively characterized the inducible expression of wild type G␣ 12 and constitutively active QL␣ 12 using the Tet-off MDCK cell culture model (12,13). The removal of dox for 72 h induces G␣ 12 expression (Fig. 1A), and the re-addition of dox suppresses G␣ 12 protein expression to background levels within 48 h (not shown, but in Refs. 12 and 13). We have previously demonstrated that expression of QL␣ 12 , but not wild type G␣ 12 leads to disruption of junctional complexes and inhibits dome formation (12). To determine whether G␣ 12 regulated apoptosis in MDCK cells, G␣ 12 -and QL␣ 12 -MDCK cells were cultured Ϯ dox for 72 h, followed by overnight serum starvation. Apoptosis was determined by propidium iodide flow cytometry (Fig. 1B), DNA laddering ( Fig. 1, C and D), and caspase 9 activity (Fig. 1E). In all assays, QL␣ 12 -MDCK (Ϫdox) cells showed a marked increase in apoptosis compared with the other conditions. In flow cytometry analysis (Fig. 1B), there was a 4 -5-fold increase in the percentage of apoptotic cells with QL␣ 12 expression when compared with ϩdox and with G␣ 12 -MDCK cells Ϯ dox. Moreover, there was a parallel reduction in the percentage of cells in the G 0 /G 1 and mitosis phases of the cell cycle. Fig. 1C shows the DNA laddering in G␣ 12 -and QL␣ 12 -MDCK cells Ϯ dox. Total intact genomic DNA was reduced to less than 50% in the QL␣ 12 expressing cells (Ϫdox) (quantified in Fig. 1D), whereas there was little effect in G␣ 12 -MDCK cells Ϯ dox. However, compared with G␣ 12 -MDCK cells, there was a small reduction in genomic DNA of QL␣ 12 -MDCK cells in ϩdox conditions (about 10 -15%). This may reflect incomplete suppression of the tetracycline responsive element leading to a small amount of QL␣ 12 expression (although QL␣ 12 protein is not detectable by Western in ϩdox, Fig. 1A). Finally, to confirm the increase in apoptosis and to determine the role of the intrinsic pathway, we assessed the activity of caspase 9. Consistent with the flow cytometry results and DNA laddering, there was a marked increase in caspase 9 activity in the QL␣ 12 expressing cells that was not seen in other conditions (Fig. 1E). This suggests that QL␣ 12 activates the intrinsic apoptotic pathway, and this is distinct from apoptosis induced by dome formation and confluence, which is associated with distal activation of caspase 8 but not caspase 9 (25).
The Effect of G␣ 12 Expression on Proliferation-The results from flow cytometry (Fig. 1B) revealed little change in the fraction of cells in the synthesis stage suggesting that proliferation may not be increased under these conditions. To address the effects of G␣ 12 on proliferation in the inducible MDCK cell model, proliferation was quantified by measuring mitochondrial dehydrogenase activity (Fig. 1F), a well established correlate of cell proliferation (27). Proliferation was measured in confluent and subconfluent monolayers. When G␣ 12 -and QL␣ 12 -MDCK cells were plated at confluence (identical to conditions for apoptosis assays) and switched to Ϯdox for 72 h, there were no significant differences in proliferation (Fig. 1F). However, when G␣ 12 -MDCK cells were plated at subconfluent density (ϳ50%) and switched into Ϯdox after plating, there was increased proliferation in G␣ 12 -and QL␣ 12 -MDCK cells (Ϫdox) at 72 h (Fig. 1G). Fig. 1G also shows that at 24 and 48 h proliferation rates were minimally stimulated in G␣ 12 -and Q␣ 12 -MDCK cells (Ϯdox) compared with Tet-off control cells.
These results indicate that G␣ 12 stimulates proliferation in MDCK cells when cultured under specific conditions. QL␣ 12 Expression Results in Reduced Bcl-2 Expression-We next examined the effects of G␣ 12 expression on the protein levels of the two major anti-apoptotic Bcl-2 family proteins, Bcl-2 and Bcl-xL. Fig. 2 (A and B) shows that inducing QL␣ 12 (Ϫdox) leads to nearly complete loss of Bcl-2 protein, in comparison with ϩdox and with G␣ 12 -MDCK cells (Ϯdox). In contrast, there was no appreciable difference in the levels of Bcl-xL protein for any of the conditions (Fig. 2A). The absence of any effect of G␣ 12 expression on Bcl-2 and Bcl-xL protein levels is consistent with the lack of an effect in G␣ 12 -MDCK cells on apoptosis. To establish whether the QL␣ 12 -stimulated loss of Bcl-2 was mediated by changes in gene expression, semi-quantitative RT-PCR was performed for each condition. Fig. 2 (C and D) reveals no significant differences in Bcl-2 transcript in the QL␣ 12 -MDCK cells in Ϫdox when compared with the other conditions.
Expression of QL␣ 12 Activates JNK1-To screen for candidate signaling molecules activated in QL␣ 12 -MDCK cells in Ϫdox, a phosphoprotein screen comparing lysates from QL␣ 12 -MDCK cells Ϯ dox was performed (Kinexus Kinetworks TM ). QL␣ 12 expression (Ϫdox) specifically increased the phosphorylation of JNK1 and completely inhibited the phosphorylation of ERK1, without significantly affecting other MAPK family members, JNK2, ERK2, or p38 (Fig. 3A). To confirm the increased JNK1 activity detected in the phosphoprotein screen, JNK1 activity was determined by phospho-JNK1 Western blot (Fig. 3, B and C) and by direct measurement of JNK activity (Fig. 3, D and E). Fig. 3 shows that inducing QL␣ 12 protein expression (Ϫdox) led to a nearly 15-fold induction of JNK1 activity as determined by p-JNK1 immunoreactivity, a value similar to the effect seen in the phosphoprotein screen. QL␣ 12 expression did not significantly affect total JNK1 levels. Additionally, Fig. 3B shows that ERK1 phosphorylation is inhibited by overexpression of both G␣ 12 and QL␣ 12 , whereas ERK2 phosphorylation remains unchanged. JNK activity was also directly measured in these cell lines during a time course of G␣ 12 protein induction (Fig. 3D). As previously described, G␣ 12 is induced at low levels by 24 h (not apparent in G␣ 12 Western blot, Fig. 3D), and is steadily expressed by 72 h. There was a small amount of baseline JNK activity in both G␣ 12 -and QL␣ 12 -MDCK cells in ϩdox (no G␣ 12 expression, Fig. 3D). Within 24 h of G␣ 12 expression, there was a significant increase in JNK activity in QL␣ 12 -MDCK cells that remained elevated through 72 h (Fig. 3, D and E). There was a small but consistent increase in JNK activity in G␣ 12 -MDCK cells at 72 h in Ϫdox (Fig. 3D). Taken together, QL␣ 12 significantly stimulates JNK activity in this MDCK cell model.
Activation of JNK1 Leads to Proteasomal Degradation of Bcl-2-Phosphorylation of Bcl-2 inactivates its anti-apoptotic properties and leads to its degradation (23). To determine FIGURE 3. QL␣ 12 stimulates JNK1 phosphorylation and induces its activation. A, phosphoprotein screen. A Kinexus TM phosphorylation screen of QL␣ 12 -MDCK cell lysates (Ϯdox for 72 h) was performed as described under "Experimental Procedures." Relative JNK1, JNK2, Erk1, ERK2, and p38 phosphorylation in the two conditions is shown. B, JNK1 and ERK1 phosphorylation. Western blots for pThr 183 /pTyr 185 -JNK1, total JNK1, pThr 202 /pTyr 204 -ERK, and total ERK were performed on G␣ 12 -and QL␣ 12 -MDCK cell lysates (Ϯdox for 72 h). The blot was stripped and reprobed for ␤-actin. C, quantification of JNK1 phosphorylation. Western blots were quantified for pThr 183 /pTyr 185 -JNK1 after normalization to total JNK1 using NIH Image. Results are the mean Ϯ S.E. of five independent experiments. D, JNK activity. JNK was pulled down from lysates of G␣ 12 and QL␣ 12 -MDCK cells (Ϯdox for 24, 48, or 72 h) using GST-cJun and incubated in vitro with 200 mM ATP to assess kinase activity as measured by GST-cJun phosphorylation. All images were obtained from the same exposure of the same Western blot. E, JNK activity quantification. Western blots were quantified for phospho-cJun after normalization with the ␤-actin using NIH Image. Results are the mean Ϯ S.E. of three independent experiments. *, significance at p Ͻ 0.05. whether JNK1 activation in QL␣ 12 -MDCK cells was important for loss of Bcl-2 expression, Bcl-2 protein levels were analyzed in G␣ 12 -and QL␣ 12 -MDCK cells in the presence and absence of the highly specific JNK inhibitor SP600125 (IC 50 ϭ 100 nM) (28). JNK1 directly phosphorylates Bcl-2 in response to microtubule damage (29), and we hypothesized that G␣ 12 -stimulated JNK activity may lead to loss of Bcl-2 expression. Fig. 4 shows that there was no effect of JNK inhibition on Bcl-2 expression in G␣ 12 -MDCK cells (Ϯdox), and in QL␣ 12 -MDCK cells ϩ dox. However, the QL␣ 12 -induced loss of Bcl-2 (Ϫdox) was nearly completely inhibited in cells treated with the JNK inhibitor. Furthermore, inhibiting Bcl-2 degradation through the proteasome pathway with lactacystin also nearly completely inhibited the QL␣ 12 -stimulated loss of Bcl-2. Taken together, these findings support the notion that G␣ 12 stimulates apoptosis by JNK activation and Bcl-2 phosphorylation leading to Bcl-2 degradation through the proteasome pathway.
QL␣ 12 Induces Expression of IB␣ through JNK1 and Bcl-2-NF-B inhibits apoptosis through numerous pathways and is directly regulated by its inhibitor, IB␣ (30). NF-B participates in a cyclic regulatory pathway with Bcl-2. NF-B stimulates Bcl-2 transcription (31,32), and Bcl-2 sustains NF-B activation by enhancing the degradation of IB␣ (summarized in Fig.  10). To determine whether NF-B signaling plays a role in G␣ 12 -stimulated apoptosis, we examined NF-B and IB␣ expression in G␣ 12 -and QL␣ 12 -MDCK cells (Ϯdox; Fig. 5). In QL␣ 12 -MDCK cells (Ϫdox), there was significant induction of IB␣ protein levels in comparison with G␣ 12 -MDCK (Ϯdox) and QL␣ 12 -MDCK (ϩdox), whereas the expression of NF-B was relatively constant (Fig. 5A). To establish whether the QL␣ 12 -stimulated induction of IB␣ was mediated through JNK1, G␣ 12 -and QL␣ 12 -MDCK cells Ϯ dox were incubated with the JNK inhibitor SP600125. Fig. 5B shows that there was no detectable IB␣ in QL␣ 12 -MDCK cells in ϩdox, whereas the QL␣ 12 -MDCK cells in Ϫdox demonstrated a large increase in IB␣ expression that was nearly completely inhibited with SP600125. As expected, inhibiting the proteasome with lactacystin had subtle effects on IB␣ expression in both G␣ 12 -and QL␣ 12 -MDCK cells Ϯ dox (Fig. 5, B and C). These findings indicate that G␣ 12 -stimulated JNK1 increases IB␣ expression and this likely contributes to G␣ 12 -mediated apoptosis.
JNK1 Activation in QL␣ 12 -MDCK Cells Is Mediated by PP2A-PP2A is a major cellular serine/threonine phosphatase and a known regulator of apoptosis (33). We recently reported that  G␣ 12 binds to the A␣ subunit of PP2A and that QL␣ 12 stimulates PP2A activity in QL␣ 12 -MDCK cells (21). To assess the potential role of PP2A in this pathway, G␣ 12 -and QL␣ 12 -MDCK cells (Ϯdox) were cultured in the presence of PP2A inhibitors fostriecin (IC 50 ϭ 1.5-5.5 nM) or okadaic acid (IC 50 ϭ 0.1-0.3 nM) (34). Fig. 6A shows the QL␣ 12 -induced (Ϫdox) increase in p-JNK1 was completely inhibited by fostriecin and okadaic acid, and the inhibitors had no demonstrable effect on total JNK1 levels in G␣ 12 -or QL␣ 12 -MDCK cells Ϯ dox. This finding was confirmed by determining JNK activity in each condition (Fig. 6B). There was a small but consistent increase in JNK activity in G␣ 12 -MDCK cells (Ϫdox) similar to what was seen in other experiments (Fig. 4, D and E). As expected, QL␣ 12 -MDCK cells (Ϫdox) demonstrated a large increase in JNK activity that was nearly completely inhibited to baseline levels with okadaic acid or fostriecin (Fig. 6, B and C). To determine whether inhibiting PP2A in MDCK-II cells leads to loss of Bcl-2 expression, Tet-off MDCK cells were cultured for 72 h in the presence of PP2A inhibitors, okadaic acid and fostriecin. As reported in other cell types, there was complete loss of Bcl-2 expression, and increased apoptosis (not shown) with PP2A inhibition (Fig. 6D), and identical findings were seen in G␣ 12and QL␣ 12 -MDCK cells Ϯ dox (results not shown). This suggests that potent inhibition of PP2A with okadaic acid and fostriecin cannot be overcome by G␣ 12 stimulation of PP2A, and other signaling pathways are likely to lead to Bcl-2 phosphorylation and degradation in the presence of PP2A inhibition.
Thrombin Stimulates Endogenous G␣ 12 Activation and Apoptosis in MDCK-II and HEK293 Cells through JNK-To determine the physiological relevance of G␣ 12 -coupled signaling in inducing apoptosis, we examined the effects of thrombin stimulation of endogenous G␣ 12 on apoptosis in two nontransfected cell lines. MDCK-II and HEK293 cells were stimulated with ␣-thrombin, a known agonist for G␣ 12 -coupled protease-activated receptor family of membrane receptors (35). Fig. 7A shows the results of the flow cytometry analysis for MDCK-II cells and HEK293 cells serum starved overnight and then stimulated with thrombin or vehicle for 24 h. Similar to the effects of QL␣ 12 expression in MDCK cells (Fig. 1D), there was a significant increase in the percentage of apoptotic cells with thrombin stimulation. In addition, the increase in apoptosis was associated with decreases in G 0 /G 1 and mitotic stages, mirroring the observations in the QL␣ 12 MDCK cells (Ϫdox). The flow cytometry analysis was repeated in the presence of thrombin and the JNK inhibitor SP600125, and the thrombin-stimulated increase in apoptosis was completely inhibited in MDCK cells, and partially (but significantly) inhibited in HEK293 cells. To confirm that thrombin stimulation of apoptosis was associated with G␣ 12 activation, GST-TPR pull-downs were performed on cell lysates obtained from MDCK-II and HEK293 cells stimulated with thrombin at specific time points. The TPR domain of protein phosphatase 5 interacts with active (GTPliganded) conformations of G␣ 12 and G␣ 13 (36,37). This assay was first validated by comparing GST pull-downs of GST and GST-TPR on MDCK cell lysates prepared from G␣ 12 -and QL␣ 12 -MDCK cells cultured in Ϯdox (Fig. 7B). Western blots of pull-downs with GST alone did not detect G␣ 12 in any condition, and there was no detectable G␣ 12 seen in GST-TPR pull-downs from G␣ 12 -MDCK cells Ϯ dox. However, there was easily detectable G␣ 12 from QL␣ 12 -MDCK cells in Ϫdox that was not apparent in GST-TPR pull-downs from QL␣ 12 -MDCK cells ϩ dox (Fig. 7B). Next, we utilized this assay to confirm G␣ 12 activation in thrombin-treated cells, and Fig. 7C shows that G␣ 12 was robustly activated within 20 min of stimulation, and G␣ 12 remained activated through 24 h. These findings reveal that thrombin stimulates apoptosis in a JNK-dependent manner and activates endogenous G␣ 12 over 24 h. Thrombin Stimulates Apoptosis in MDCK and HEK293 Cells through Similar Pathways-To determine whether the same apoptosis pathway(s) identified in QL␣ 12 -MDCK cells were activated in non-transfected cells, a time course of thrombin effects on Bcl-2, phospho-JNK1, pSer 70 -Bcl-2, IB␣, and NF-B were determined at specific time points in MDCK-II and HEK293 cells (Fig. 8). Analogous to the QL␣ 12 -simulated JNK1 activation and loss of Bcl-2, thrombin had a similar effect in both cell types. Bcl-2 phosphorylation increased within the first 4 h and Bcl-2 protein levels steadily declined in both cell types to less than 25% at 24 h. JNK1 was phosphorylated within 20 min, and peaked at about 1 h without affecting JNK levels. JNK1 phosphorylation preceded Bcl-2 phosphorylation in both cell types. Similarly, IB␣ expression was induced at the early time points and increased over the subsequent 24 h although there was little effect on NF-B levels. Although the time course of thrombin-stimulated activation of JNK and loss of Bcl-2 differed slightly in MDCK and HEK293 cells, the overall findings were remarkably similar. To confirm JNK activation and regulation of Bcl-2 expression, thrombin-stimulated MDCK and HEK293 cells were analyzed at 24 h ϮJNK inhibition with SP600125 (Fig. 8, E-H). The degradation of Bcl-2 was prevented in both cell types in the presence of the JNK inhibitor (Fig. 8, E and F). When normalized to total Bcl-2 levels, the phospho-Bcl-2 levels were increased in both cell types with thrombin stimulation (consistent with targeting to degradation), and this was blocked in the presence of JNK inhibition with SP600125 (Fig. 8H). The time course of activation for these pathways correlates with the timing of G␣ 12 activation, but we cannot exclude the possibility that thrombin also activates additional pathways that contribute to the stimulation of apoptosis in these cells.
Thrombin/G␣ 12 -regulated Apoptosis Requires PP2A-To establish that thrombin-stimulated G␣ 12 leads to loss of Bcl-2 and JNK activation in a PP2A-dependent manner, HEK293 cells were characterized in the presence and absence of silenced PP2A. The catalytic subunit was silenced, and Fig. 9A shows a 70% reduction in catalytic subunit expression by Western blot. In the absence of thrombin, silencing PP2A or the use of control oligonucleotides did not affect apoptosis (Fig. 9B). However, as expected with thrombin stimulation, there was a significant increase in apoptosis that was not affected by control oligonucleotides or the transfection procedure (Fig. 9C). In cells with silenced PP2A catalytic subunit, there was partial rescue of the increased apoptosis. This partial reduction in thrombin-stimulated apoptosis in PP2A-silenced HEK293 cells was significant, similar to the effect seen with JNK inhibition (Fig. 7A, and expected in an experiment where 30 -50% of the cells were silenced). To confirm that this pathway regulates JNK and Bcl-2 expression, HEK293 cells Ϯ silenced PP2A were analyzed by Western blot. Fig. 9D shows that thrombin-stimulated HEK293 cells with silenced PP2A do not exhibit increased phospho-JNK or loss of Bcl-2 expression as seen in the non-silenced thrombin-stimulated conditions.

DISCUSSION
Cell proliferation and apoptosis are tightly regulated in the normal state, and disturbances in either process lead to major pathophysiologic changes resulting in numerous diseases. G proteins regulate both proliferation and apoptosis, but the mechanisms maintaining the appropriate balance are not well defined. Furthermore, the specific mechanisms regulating proliferation and apoptosis depend upon the stimulus and cell type. The findings of this study identify an important role for G␣ 12 in regulating epithelial cell apoptosis. To our knowledge, this is the first definitive demonstration that G␣ 12 regulates apoptosis and that this pathway can be activated through endogenous signaling pathways. Furthermore, these studies reveal that G␣ 12 stimulates apoptosis through well established pathways that include Bcl-2 degradation, signaling through Jun kinases, and up-regulation of the NF-B inhibitor, IB␣. In addition, these findings support our previous finding that G␣ 12 regulates PP2A, and reveal both G␣ 12dependent and -independent regulation of apoptosis through PP2A. These observations are summarized in Fig. 10.
The studies on G protein regulation of apoptosis in specialized cells implicate several pathways depending upon the cell type. G␣ protein overexpression in neuronal cells increased sensitivity to apoptotic stimuli that was independent of cAMP and PKA activity (9). In renal tubular epithelium, tumor necrosis factor ␣-induced apoptosis that was associated with activation of NF-B through a pertussis toxin-sensitive G protein pathway (4). In the heart, muscarinic receptor activation of G␣ q pathways protected cells from apoptosis (7), through increased Akt activity, but in HeLa cells, G␣ q stimulated apoptosis by decreasing Akt activation utilizing a Rho-dependent pathway FIGURE 7. Thrombin-stimulated activation of G␣ 12 induces JNK-dependent apoptosis in MDCK-II and HEK293 cells. A, flow cytometry. Tet-Off MDCK-II cells and HEK293 cells were cultured to confluence, serum starved overnight, and stimulated with thrombin (2 units/ml) for 24 h. Subsequently they were fixed and stained with propidium iodide and analyzed by flow cytometry. Cells were quantified according to cell cycle phase. Results are the mean Ϯ S.E. of three independent experiments. B, validation of GST-TPR assay. Activated G␣ 12 was pulled down from lysates of G␣ 12 -and QL␣ 12 -MDCK cells (Ϯdox for 72 h) using GST-TPR or GST as described under "Experimental Procedures." C, MDCK-II cells (non-transfected) were cultured to confluence, serum starved overnight, and stimulated with thrombin for 20 min, 1 h, 4 h, 8 h, or 24 h and activated G␣ 12 was pulleddown from ϳ800 g of cell lysate using GST-TPR. 5% of the input used for the pull down was loaded in the lysate lanes (ϳ40 g). (8). In several hormone-dependent cancers, gonadotropin-releasing hormone antagonists inhibited apoptosis through direct effects on G␣ i pathways (6). G␣ 12 and G␣ 13 have been implicated in many cellular processes.
Whereas several studies have implicated G␣ 12 in proliferation and neoplasia (38,39), recent studies suggest that G␣ 12 does not stimulate proliferation in epithelial tumors, but promotes cell migration. Expression of G␣ 12 was elevated in breast and prostate cancer, and thrombin stimulation resulted in increased cell invasion but not proliferation (40,41). However, the effects of up-regulated G␣ 12 levels on apoptosis were not addressed in these studies. Interestingly, we also found no significant difference in proliferation in overexpressing G␣ 12 or QL␣ 12 -MDCK when cultured at confluence, but did detect increased proliferation after 72 h in non-confluent monolayers. These findings suggest that the mitogenic potential of increased G␣ 12 expression or activation of G␣ 12 depends upon the cell type and may be influenced by cellular confluence. Remarkably, there have been few studies on apoptosis regulated by the G␣ 12 /G␣ 13 family. There is only one report on the role of G␣ 12 in apoptosis, and this was in a transient transfection system (18). Additionally, G␣ 13 stimulated apoptosis in Rat 1A cells through JNK (19), an important MAPK regulating apoptosis in other signaling pathways. Based upon these observations, we utilized previously characterized MDCK cells with inducible G␣ 12 and QL␣ 12 to ask whether G␣ 12 regulates apoptosis in the MDCK cell model epithelia. The inducible expression of QL␣ 12 led to increased apoptosis as assessed by flow cytometry, DNA laddering, and caspase 9 activity. Unlike proliferation that was stimulated by overexpressing G␣ 12 (at 72 h), there was no significant effect of G␣ 12 overexpression on apoptosis.
The mechanism of G␣ 12 -stimulated apoptosis in MDCK cells involves at least two well established apoptosis signaling pathways. At the center of these pathways is the ratio of pro-apoptotic and anti-apoptotic Bcl family of proteins. The loss of the anti-apoptotic protein Bcl-2 permits the homodimerization of the pro- apoptotic protein Bax and the stimulation of apoptosis. For example, osteoblast and osteoclast apoptosis has been associated with an increase in Bax/Bcl-2 ratio (42), and apoptosis of renal proximal tubular epithelial cells due to local hydrogen peroxide generation is also associated with this shift in Bax/ Bcl-2 ratio (43). Likewise, the mechanism of G␣ 12 -stimulated apoptosis presented in our studies involves degradation of Bcl-2 that is mediated by JNK stimulation. JNK has been implicated in several G␣ 12 -initiated signaling cascades, and the JNK family regulate several apoptosis pathways (44). JNK localizes at the mitochondrial membrane, and activated JNK1 has been shown to directly phosphorylate Bcl-2 in response to microtubule damage (29). The decrease in the percentage of QL␣ 12 expressing MDCK cells in the mitotic phase but not the synthesis phase (Fig. 1B) is also consistent with JNK phosphorylation of Bcl-2 at the G 2 /M checkpoint (45). It is possible that G␣ 12 activation of JNK1 specifically causes the targeting of Bcl-2 to degradation at this checkpoint, causing these cells to undergo apoptosis rather than continue into mitosis.
Selective activation of JNK1 but not JNK2 was associated with renal fibrosis and tubular cell apoptosis (46), phenotypes linked to polycystic kidney disease (47). Furthermore, polycystin-1, the major protein mutated in polycystic kidney disease may be a novel G protein-coupled receptor (48,49) and interacts with G␣ 12 (50). Polycystin-1 activates JNK and the associated AP-1 transcription factors by coupling to various heterotrimeric G proteins including G␣ 12 (49,51). Additionally, Bcl-2 knock-out mice develop a cystic phenotype (52). Loss of Bcl-2 may cause cystogenesis, and consistent with this possibility, overexpression of Bcl-2 in MDCK cells prevented cyst formation associated with anoikis-induced apoptosis (53). In vivo, the JNK-specific inhibitor SP600125 blocked Bcl-2 phosphorylation and inhibited apoptosis associated with brain ischemia (54).
In QL␣ 12 expressing MDCK cells, apoptosis is clearly stimulated. However, the in vivo relevance of these observations needed to be established in non-transfected cells. Therefore, we FIGURE 9. Knock-down of PP2A C␣ inhibits thrombin-stimulated apoptosis by preventing JNK1 activation and Bcl-2 degradation. A, HEK293 cells were transfected with Lipofectamine alone (LFA), scrambled (scr) siRNA, or a mixture of silencing oligonucleotides for the PP2A catalytic subunit (C␣) as described under "Experimental Procedures." PP2A C subunit and Bcl-2 expression were assessed after 48 h by Western blot. Relative expression of C␣ compared with non-transfected HEK cells is shown below and normalized to the ␤-actin expression. B and C, HEK293 cells were transfected and cultured as described above, and serum starved overnight. Cells were then left untreated (B) or stimulated with thrombin (Thr) for 24 h (C) and analyzed by flow cytometry as described under "Experimental Procedures." Results are the mean Ϯ S.E. of three independent experiments. D, Western blot of pThr 183 /pTyr 185 -JNK1, total JNK, and Bcl-2 expression in HEK293 cells that were non-transfected (control) or transfected with scrambled or specific oligonucleotides for silencing PP2A catalytic subunit and then thrombin stimulated for 8 h. *, significance at p Ͻ 0.05; **, significance at p Ͻ 0.05 when compared with *. FIGURE 10. A schematic representation of the G␣ 12 -stimulated apoptosis signaling pathways. Activated G␣ 12 (QL␣ 12 expression or by thrombin stimulation) stimulates PP2A, resulting in the increased phosphorylation of JNK1 through the suppression of ERK1. Activated JNK1 phosphorylates Bcl-2 leading to Bcl-2 degradation. The absence of Bcl-2 results in loss of the heterodimerization with BAX, permitting BAX homodimerization with activation of the intrinsic apoptosis cascade through caspases-9 and -3. In addition, there are G␣ 12 -independent effects of PP2A on Bcl-2 phosphorylation (dashed lines) and other pathways from G␣ 12 to JNK that are PP2A independent (double-headed arrows). Furthermore, G␣ 12 -stimulated JNK activity and degradation of Bcl-2 leads to up-regulation of ⌱B␣ protein levels. Increased ⌱B␣ protein expression results in reduced NF-B activity and stimulation of apoptosis. Arrows indicate stimulation. Lines ending in a dash indicate inhibition.
utilized MDCK-II and HEK293 cells to validate the observations from the G␣ 12 -MDCK cell culture model. Thrombin activates G␣ 12 -coupled protease-activated receptors (35) and has been implicated in numerous disease phenotypes, including ischemia-reperfusion injury (54). Thrombin significantly stimulated apoptosis in both MDCK-II and HEK293 cells. To confirm that thrombin activated endogenous G␣ 12 , we first validated the GST-TPR pull-down assay-activated G␣ 12 (described in Refs. 36 and 37). Utilizing this novel assay, we confirmed that thrombin activated endogenous G␣ 12 throughout the 24 h of thrombin exposure. In addition, thrombin activated JNK and led to loss of Bcl-2 in MDCK-II and HEK293 cells, similar to what was observed in the inducible QL␣ 12 -MDCK model. Additional studies support the notion that thrombin is important in regulating apoptosis. Recently, it was shown that intranigral injection of thrombin stimulated apoptosis by increasing JNK activity and dramatically reducing Bcl-2 expression, leading to nigral dopaminergic neurodegeneration (55), a pathology of Parkinsonism. However, our studies were not designed to distinguish between thrombin-stimulated G␣ 12 and other potential pathways important to apoptosis.
Like G proteins, PP2A has many cellular functions. G␣ 12 binds PP2A and stimulates phosphatase activity (20), and the binding domains were recently identified (21). The current results extend these observations and reveal regulation of apoptosis through PP2A using at least 2 different pathways: G␣ 12dependent activation of JNK and a G␣ 12 -independent pathway. The answer to the question of why G␣ 12 -stimulated PP2A activity does not offset JNK activation and directly inhibit Bcl-2 degradation may reside in the relative potency of these pathways. The stoichiometry, localization of signaling components, and specific stimuli may affect which pathway predominates. The differing effects on apoptosis when PP2A is partially silenced versus chemically inhibited is likely to be explained by such a mechanism. The partial silencing of PP2A prevents G␣ 12dependent effects, but does not alter direct effects of PP2A on Bcl-2 expression. This suggests that the available pool of functional PP2A contributes to determining which pathways are activated. In addition, PP2A is highly abundant, whereas the relative expression of G␣ 12 is much lower. Therefore, the finding of G␣ 12 -dependent and -independent effects of PP2A on apoptosis is not surprising (see Fig. 10). Although G␣ 12 has been previously shown to activate JNK through numerous other signaling pathways, including small G proteins (56), Src (57), and various MAPK kinases (58), this is the first report linking PP2A to JNK activation by G␣ 12 . In intestinal epithelial cells, PP2A stimulated apoptosis through a JNK-dependent mechanism that required ERK1 inactivation (59). The mechanism of PP2A activation of JNK in our studies was not explicitly examined, but we also observed loss of ERK1 phosphorylation (Fig. 3, A and B). This leads us to speculate that ERK1 inhibition may be released by G␣ 12 -PP2A activation, and this contributes to JNK activation. The role of PP2A and JNK stimulating apoptosis in intestinal epithelia in combination with our findings in MDCK cells suggests that G␣ 12 -coupled signaling pathways are a general mechanism regulating apoptosis in epithelia. Additional studies will be necessary to definitively answer this question, but the coupling of G␣ 12 to numerous receptors regulating proliferation and apoptosis make this an intriguing possibility (and potential therapeutic target).
Another major mechanism regulating apoptosis involves NF-B. NF-B prevents apoptosis by regulating transcription of pro-apoptotic genes (reviewed in Ref. 60). G proteins also regulate apoptosis through NF-B. For example, G␣ q stimulates NF-B activity through the phosphatidylinositol 3-kinase pathway (61), and inhibits NF-B activity by preventing IB␣ degradation (62). In addition, G␣ 13 stimulates NF-B activity through reactive oxygen species and Ca 2ϩ -mediated signaling pathways (63). Recently, G␣ 12 -mediated activation of JNK was reported to induce proteasomal degradation of IB␣, leading to the activation of NF-B in the regulation of cyclooxygenase transcription. However, this study did not examine apoptosis (64). Here, we reveal novel suppression of NF-B activity by G␣ 12 through the targeted up-regulation of IB␣ expression. In the absence of changes in Bcl-2 gene expression, it is likely that Bcl-2 is upstream of the NF-B pathway in this model (see Fig.  10). Our finding of enhanced Bcl-2 degradation is also consistent with the up-regulation of IB␣, although the relative role of the NF-B pathway in G␣ 12 -stimulated apoptosis remains to be elucidated.
In conclusion, these studies reveal that G␣ 12 stimulates apoptosis in epithelial cells, and the mechanisms are nearly identical with inducible expression of QL␣ 12 and with thrombin activation of endogenous G␣ 12 . Activation of G␣ 12 leads to PP2A-mediated JNK1 activation, proteasomal degradation of Bcl-2, and stimulation of apoptosis (see Fig. 10). This identifies novel roles for PP2A and JNK1 in G␣ 12 signaling and specifically in the regulation of apoptosis. Furthermore, G␣ 12 -stimulated JNK activity also fundamentally regulates NF-B by stimulating expression of its inhibitor IB␣. The numerous reports of apoptosis regulation through thrombin and identification of pathways that utilize JNK, PP2A, and Bcl-2 suggest a general mechanism with proximal regulation by G␣ 12 . Future studies will target G␣ 12mediated apoptosis in vivo to develop therapies for specific diseases where these signaling pathways have gone awry.