Constitutively Active Gα16 Stimulates STAT3 via a c-Src/JAK- and ERK-dependent Mechanism*

The hematopoietic-specific Gα16 protein has recently been shown to mediate receptor-induced activation of the signal transducer and activator of transcription 3 (STAT3). In the present study, we have delineated the mechanism by which Gα16 stimulates STAT3 in human embryonic kidney 293 cells. A constitutively active Gα16 mutant, Gα16QL, stimulated STAT3-dependent luciferase activity as well as the phosphorylation of STAT3 at both Tyr705 and Ser727. Gα16QL-induced STAT3 activation was enhanced by overexpression of extracellular signal-regulated kinase 1 (ERK1), but was inhibited by U0126, a Raf-1 inhibitor, and coexpression of the dominant negative mutants of Ras and Rac1. Inhibition of phospholipase Cβ, protein kinase C, and calmodulin-dependent kinase II by their respective inhibitors also suppressed Gα16QL-induced STAT3 activation. The involvement of tyrosine kinases such as c-Src and Janus kinase 2 and 3 (JAK2 and JAK3) in Gα16QL-induced activation of STAT3 was illustrated by the combined use of selective inhibitors and dominant negative mutants. In contrast, c-Jun N-terminal kinase, p38 MAPK, RhoA, Cdc42, phosphatidylinositol 3-kinase, and the epidermal growth factor receptor did not appear to be required. Similar observations were obtained with human erythroleukemia cells, where STAT3 phosphorylation was stimulated by C5a in a PTX-insensitive manner. Collectively, these results highlight the important regulatory roles of the Ras/Raf/MEK/ERK and c-Src/JAK pathways on the stimulation of STAT3 by activated Gα16. Demonstration of the involvement of different kinases in Gα16QL-induced STAT3 activation supports the involvement of multiple signaling pathways in the regulation of transcription by G proteins.

The hematopoietic-specific G␣ 16 protein has recently been shown to mediate receptor-induced activation of the signal transducer and activator of transcription 3 (STAT3). In the present study, we have delineated the mechanism by which G␣ 16 stimulates STAT3 in human embryonic kidney 293 cells. A constitutively active G␣ 16 mutant, G␣ 16 QL, stimulated STAT3-dependent luciferase activity as well as the phosphorylation of STAT3 at both Tyr 705 and Ser 727 . G␣ 16 QL-induced STAT3 activation was enhanced by overexpression of extracellular signal-regulated kinase 1 (ERK1), but was inhibited by U0126, a Raf-1 inhibitor, and coexpression of the dominant negative mutants of Ras and Rac1. Inhibition of phospholipase C␤, protein kinase C, and calmodulin-dependent kinase II by their respective inhibitors also suppressed G␣ 16 QL-induced STAT3 activation. The involvement of tyrosine kinases such as c-Src and Janus kinase 2 and 3 (JAK2 and JAK3) in G␣ 16 QL-induced activation of STAT3 was illustrated by the combined use of selective inhibitors and dominant negative mutants. In contrast, c-Jun N-terminal kinase, p38 MAPK, RhoA, Cdc42, phosphatidylinositol 3-kinase, and the epidermal growth factor receptor did not appear to be required. Similar observations were obtained with human erythroleukemia cells, where STAT3 phosphorylation was stimulated by C5a in a PTX-insensitive manner. Collectively, these results highlight the important regulatory roles of the Ras/Raf/MEK/ERK and c-Src/JAK pathways on the stimulation of STAT3 by activated G␣ 16 . Demonstration of the involvement of different kinases in G␣ 16 QL-induced STAT3 activation supports the involvement of multiple signaling pathways in the regulation of transcription by G proteins.
In addition to their classical roles as second messenger regulators, heterotrimeric G proteins have been implicated as mitogenic signal transmitters. The discovery of activating G protein mutations in various disease states highlights their roles in normal and aberrant growth (1). To date, a number of G␣ subunits have been shown to stimulate mitogenesis and induce neoplastic growth via initiation of intracellular signal-ing cascades that lead to the activation of mitogen-activated protein kinases (MAPKs, 1 Refs. 2 and 3). In addition to MAPKs, other critical molecules such as signal transducers and activators of transcription (STATs) have also been shown to participate in the transduction of proliferative signals (4).
STATs are latent cytoplasmic transcription factors that transduce signals from the cell membrane to the nucleus upon tyrosine phosphorylation (5). They were first identified as mediators of cellular responses to cytokines (6), but later it became apparent that they are also involved in mitogenic growth factor signaling (7). Binding of cytokines or growth factors to their cognate receptors leads to receptor dimerization and activation of receptor-associated Janus kinases (JAKs), resulting in the recruitment and homo-or heterodimerization of STAT proteins. Activated STAT proteins are then translocated to the nucleus to regulate gene expression. STAT activation by other non-receptor tyrosine kinases has also been demonstrated. Transformation of mammalian fibroblasts by viral Src (v-Src) specifically induces constitutive activation of STAT3 (8). Cellular Src (c-Src) tyrosine kinase is involved in the activation of both STAT1 and STAT3 in platelet-derived growth factor (PDGF)-stimulated NIH-3T3 cells (9). Additionally, it has recently been demonstrated that MAPKs can phosphorylate Ser 727 on STAT3 to modulate its transcriptional activity (10), while activation of p38 MAPK and c-Jun N-terminal kinase (JNK) is thought to be required for v-Src activation of STAT3 (11).
Although activation of STAT proteins has generally been associated with cytokine and mitogenic growth factor signaling, ligands acting on G protein-coupled receptors (GPCRs) can also activate STAT proteins. Angiotensin II has been shown to induce c-Src-dependent tyrosine (Tyr 705 ) phosphorylation of STAT3 via activation of the G protein-coupled AT 1 receptor in vascular smooth muscle cells (12). ␣-Melanocyte-stimulating hormone, which enhances cellular proliferation, has been found to activate JAK2 and STAT1 in B-lymphocytes via stimulation of the melanocortin 5 receptor (13). Likewise, activation of ␣ 1 -adrenoceptors and protease-activated receptor 1 has been shown to induce tyrosine phosphorylation of JAK2, Tyk2, and STAT1 in vascular smooth muscle cells (14).
While activation of STATs in response to GPCR stimulation has been reported, the involvement of STATs in G␣-mediated transformation of cells is beginning to emerge (15). Expression of constitutively active G␣ o in NIH-3T3 cells results in Src-de-pendent activation of STAT3, which leads to cellular transformation. Similarly, expression of constitutively active G␣ i2 in NIH-3T3 cells increases STAT3 activity (4). Conversely, expression of a dominant negative mutant of G␣ i2 inhibits Src kinase activity and Tyr 705 phosphorylation of STAT3, leading to a reduction of v-fms-induced proliferation in NIH-3T3 cells (16). These data suggest that the STAT3 pathway may play a vital role in the G␣ subunit regulation of cell proliferation and transformation.
With the recent demonstration of STAT3 involvement in G␣ o -and G␣ i2 -induced cell transformation, it is reasonable to deduce that other G␣ subunits may also regulate mitogenesis via STAT3 activation. G␣ 16 , being unique in its restricted expression in hematopoietic cells (17), is also expressed in poorly differentiated leukemia cells, suggesting an association with hematopoietic cell growth and differentiation. Expression of a constitutively active G␣ 16 mutant has been shown to induce cell differentiation in rat pheochromocytoma PC12 (3) and aortic vascular smooth muscle cells, although the same mutant was found to inhibit cell growth in Swiss 3T3 cells (18). Such observations suggest that G␣ 16 may regulate cell growth and differentiation via activation of cell type-specific signal transduction pathways.
As a promiscuous G protein (19), G␣ 16 possesses the ability to link a variety of GPCRs to the regulation of MAPKs. Recently, G␣ 16 has been shown to activate JNK (3,20). Interestingly, in addition to its ability to phosphorylate c-Jun, JNK can also phosphorylate STAT3 at Ser 727 (21). It is therefore plausible that activated G␣ 16 can influence cell differentiation via MAPK-induced STAT3 signaling. Based on the exclusivity of G␣ 16 expression in hematopoietic cells and the involvement of STAT pathways in both normal and perturbed hematopoiesis (22), phosphorylation of STAT3 via G␣ 16 activation may represent an important pathway for cell differentiation and development in the immune system. Indeed, we have recently shown that G␣ 16 can support receptor-mediated activation of STAT3 in human embryonic kidney 293 (HEK 293) cells (23). In the present study, we examined the mechanism by which G␣ 16 16 , Ras, Rac1, and JNK1 cDNAs were as described previously (20,25). The luciferase reporter genes, pSTAT3-TA-luc, pGAS-TA-luc and pISRE-TA-luc, were obtained from Clontech laboratories, Inc. (Palo Alto, CA). The luciferase substrate and its lysis buffer were purchased from Roche Diagnostics (Mannheim, Germany). All antibodies were obtained from Cell Signaling and kinase inhibitors were from Calbiochem (Darmstadt, Germany). C5a was purchased from Sigma Aldrich.
Cell Culture and Transfection-HEL cells were maintained at 5% CO 2 , 37°C in RPMI 1640 with 10% fetal bovine serum, 50 units/ml penicillin and 50 l/ml streptomycin. HEK 293 cells were maintained at 5% CO 2 , 37°C in Eagle's minimum essential medium (growth medium) with 10% fetal bovine serum, 50 units/ml penicillin and 50 l/ml streptomycin. HEK 293 cells were seeded on 96-well microtiter plates at a density of 15,000 cells/well and were cultured in the growth medium at 18 to 24 h prior to transfection. They were co-transfected with various cDNAs using LipofectAMINE PLUS reagents. The transfection mixtures using 100 l/well of serum and antibiotics free OPTI-MEM medium contained 10 ng of G proteins, small GTPases or the control vector cDNAs, 0.1 g of pSTAT3-TA-luc and 0.2 l of both PLUS and Lipo-fectAMINE reagents. After 3 h of transfection, 50 l of OPTI-MEM medium containing 30% fetal bovine serum was added into the wells and incubated for another 30 h.
Luciferase Assay-30 h after transfection, cells were serum-starved for 24 h. After removal of the medium, 25 l of lysis buffer from Roche Diagnostics luciferase assay kit was added to the wells and then gently shaken on ice for 30 min. For detection, cell lysates in 25 l of lysis buffer and 25 l of luciferase substrate were measured by a microtiter plate luminometer MicroLumatPlus LB96V from EG&G Berthold.
Western Blot Analysis-HEK 293 cells were transferred on 6-well plates at a density of 5 ϫ 10 5 cells/well and were kept in the growth medium the day before transfection. They were co-transfected with various cDNAs using LipofectAMINE PLUS reagents following the supplier's instructions. After 48 h of transfection, the transfected cells were serum starved or treated with different kinase inhibitors overnight. The cells were lysed in 150 l of lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 5 mM EDTA, 40 mM NaP 2 O 7 , 1% Triton X-100, 1 mM dithiothreitol, 200 M Na 3 VO 4 , 100 M phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, 4 g/ml aprotinin, and 0.7 g/ml pepstatin) and then gently shaken on ice for 30 min. Supernatants were collected by centrifugation at 16,000 ϫ g for 5 min. HEL cells were seeded at a density of 1 ϫ 10 6 cells using serum-free medium with or without PTX (100 ng/ml) treatment overnight. Cells were treated with different kinase inhibitors for 30 min and then incubated with 100 nM C5a at 37°C for 20 min. After that, cells were lysed by lysis buffer. Proteins from the cell lysates were resolved by 12% SDS-polyacrylamide gel electrophoresis, and then transferred to Osmonics nitrocellulose membrane (Westborough, MA). Phospho-STAT3-Tyr 705 , phospho-STAT3-Ser 727 , STAT3, phospho-p38 MAPK, p38 MAPK, phospho-JNK, JNK, phospho-ERK, and ERK were detected by specific primary antibodies and horseradish peroxidase-conjugated secondary antibodies. The immunoblots were visualized by chemiluminescence with the ECL kit from Amersham Biosciences. 16 QL in HEK 293 Cells-It has previously been shown that expression of constitutively active G␣ o and G␣ i2 in NIH-3T3 cells induced STAT3 activation (4,16). In the present study, we sought to investigate the ability of constitutively active G␣ 16 to stimulate STAT3 phosphorylation and activation. Mutation of glutamine 212 to leucine (Q212L) in the conserved GTP/GDP binding domain of G␣ 16 inhibits its intrinsic GTPase activity and results in constitutive activation of G␣ 16 (3,25). To investigate the effect of G␣ 16 QL mutant on the phosphorylation state of STAT3, HEK 293 cells were transiently transfected with cDNAs encoding pcDNA1 (vector control), G␣ 16 or G␣ 16 QL. Total cell lysates prepared from the transfected cells were probed with anti-STAT3, anti-phospho-STAT3-Tyr 705 and antiphospho-STAT3-Ser 727 antisera. Expression of either G␣ 16 or G␣ 16 QL in HEK 293 cells did not affect the expression of total STAT3 as compared with the vector control (Fig. 1A). Further studies using anti-phospho-STAT3-Tyr 705 and anti-phospho-STAT3-Ser 727 antisera revealed basal levels of Ser 727 phosphorylation of STAT3 protein in both the vector control and G␣ 16expressing cells, while there was little or no Tyr 705 phosphorylation of STAT3 in these cells. In contrast, expres-sion of G␣ 16 QL led to a striking increase in both Tyr 705 and Ser 727 phosphorylation of STAT3 protein (Fig. 1A). The observed G␣ 16 QL-induced phosphorylation of STAT3 is likely to induce STAT3 transcriptional activity, as demonstrated with G␣ o QL-induced phosphorylation and activation of STAT3 (4). In order to verify a similar correlation between G␣ 16 QL-induced STAT3 phosphorylation and modulation of STAT3 transcriptional activity, we performed reporter gene assays using a pSTAT3-TA-luc construct in combination with various cDNAs (pcDNA1, G␣ 16 and G␣ 16 QL). As indicated in Fig. 1B, the magnitude of STAT3-mediated gene expression was unaffected by control vector or G␣ 16 expression. On the contrary, expression of G␣ 16 QL induced an increased level of STAT3 activation as evidenced by a significant elevation of reporter gene expression. These results suggest that activation of G␣ 16 can indeed lead to the phosphorylation of STAT3 at both Tyr 705 and Ser 727 as well as the induction of STAT3 transcriptional activity. Using two other reporter gene constructs (pGAS-TA-luc and pISRE-TA-luc), we examined the ability of G␣ 16 QL to similarly activate STAT1 and STAT1/2 transcriptional activities. The immunoblots shown represent one of three sets of immunoblots; two other sets yielded similar results. For luciferase reporter gene assays, cell lysates were used to measure luciferase activities. The luminescence emitted was quantified as the relative luminescent unit (RLU). The RLUs were expressed as a percentage of the pcDNA1 control. Data shown represent the mean Ϯ S.E. from four separate experiments performed in triplicates. *, G␣ 16 QL-induced STAT activity was significantly higher than the corresponding pcDNA1 control (Dunnett t test, p Ͻ 0.05). #, G␣ 16 QL-induced STAT3 activity was significantly inhibited as compared with those obtained with STAT3-overexpressing cells (Dunnett t test, p Ͻ 0.05). STAT1-mediated gene expression was significantly elevated in G␣ 16 QL-expressing cells, but not in G␣ 16 or vector-transfected cells (Fig. 1B). In contrast, STAT1/2-mediated gene expression was not enhanced in G␣ 16 QL-expressing cells (Fig. 1B).

Activation of STAT3 by Constitutively Active G␣
The specificity of G␣ 16 QL-induced STAT3 activation was further characterized by the dose-dependent relationship between the amount of G␣ 16 QL expression and STAT3 stimulation. As shown in Fig. 1C, the magnitude of STAT3-dependent luciferase activity was positively correlated to the concentration of G␣ 16 QL cDNA transfected into the HEK 293 cells, whereas no correlation with activity could be observed for the cDNA of G␣ 16 . The progressive increase in STAT3-dependent luciferase activity was eminent at the concentration range of 3-30 ng/ml of G␣ 16 QL cDNA. Increasing the G␣ 16 QL cDNA concentration beyond 30 ng/ml failed to induce further increase in luciferase activity, presumably because G␣ 16 QL was no longer a limiting factor. The observed saturation of STAT3-dependent luciferase activity was likely to be caused by the limited amount of pSTAT3-TA-luc available in the transfected cells.
As both Tyr 705 and Ser 727 of STAT3 appeared to be phosphorylated in G␣ 16 QL-expressing cells, we examined whether these modifications can occur independently. Using the reporter gene assay, the ability of G␣ 16 QL to induce the phosphorylation of wild-type and phosphorylation-resistant mutants of STAT3 (STAT3Y705F and STAT3S727A) was examined in HEK 293 cells. Coexpression of STAT3 and G␣ 16 QL resulted in an elevation of luciferase activity that was significantly higher than those obtained with G␣ 16 and vector control, and the magnitude of the response was identical to that of cells lacking recombinant STAT3 (Fig. 1D). In cells coexpressing the Y705F mutant of STAT3, although G␣ 16 QL was still capable of inducing STAT3-dependent luciferase activity beyond the controls, such stimulation was significantly attenuated as compared with cells coexpressing wild-type STAT3 (Fig. 1D). Similar results were obtained with the S727A mutant of STAT3. Overexpression of STAT3 and its mutants was detected with anti-STAT3 antiserum (Fig. 1D). Lack of phosphorylation of the Y705F and S727A mutants of STAT3 at Tyr 705 and Ser 727 , respectively, was confirmed with anti-phospho-STAT3 antisera (Fig. 1D). Basal as well as G␣ 16 QL-dependent Tyr 705 phosphorylation were prominently observed in STAT3-and S727A-mutant expressing cells, whereas a reduction in Tyr 705 phosphorylation was seen with the Y705F mutant. Conversely, elevations in the levels of Ser 727 phosphorylation were detected in cells expressing the wild-type and Y705F mutant of STAT3, while little or no phosphorylation at this site was detectable in S727A-expressing cells. This indicated that G␣ 16 QL-induced phosphorylation of STAT3 at Tyr 705 and Ser 727 could occur independently.
Effects of MAPKs on STAT3 Phosphorylation-The MAPK pathway has been shown to play an important role in the regulation of STAT3 signaling. Both ERKs and JNK1 have been shown to phosphorylate STAT3 at Ser 727 (10,21). Likewise, p38 MAPK has been shown to play a key role in the Ser 727 phosphorylation of STAT3 (11). Since G␣ 16 has been shown to activate JNK (4, 24), we asked if it can similarly stimulate ERK and p38 MAPK. Anti-phospho-MAPK antisera were used to probe the activities of MAPKs in G␣ 16 QL-transfectants. The activities of all three branches of MAPKs were stimulated by G␣ 16 QL, but not G␣ 16 (Fig. 2A). Next, we investigated if one or more of these MAPKs were required for G␣ 16 QL-induced STAT3 activation. Firstly, we examined the effect of MAPKs expression on STAT3 activity. As illustrated in Fig. 2B, overexpression of ERK1 significantly stimulated STAT3 activation. In contrast, overexpression of JNK1 or p38 MAPK had no effect on STAT3 stimulation, suggesting that these two pathways did not play a significant role in STAT3 regulation in HEK 293 cells.
We used a panel of kinase inhibitors to further confirm the functional role of various MAPKs on G␣ 16 QL-induced STAT3 phosphorylation and activation. The MAPK/ERK kinase 1/2 (MEK1/2) inhibitor, U0126 (10 M), inhibited G␣ 16 QL-induced STAT3 activation whereas the p38 MAPK/JNK inhibitors, SB202190 (10 M) and SB203580 (10 M), did not affect STAT3dependent luciferase activity (Fig. 2C). The ineffectiveness of both SB202190 and SB203580 further substantiated the lack of involvement of p38 MAPK and JNK in STAT3 activation. The inhibitory effect of U0126 reflects a possible involvement of MEK1/2 in STAT3 regulation. Inhibition of the upstream activator of MEK1/2, Raf-1, by its specific inhibitor (10 M) also suppressed the G␣ 16 QL-induced STAT3 activation (Fig. 2C). These observations provided additional evidence to support the regulatory role of ERK on STAT3 activation. The effects of the various inhibitors on the state of phosphorylation of STAT3 were examined with anti-phospho-STAT3 antisera. Both U0126 and the Raf-1 inhibitor, but not SB202190 and SB203580, attenuated the G␣ 16 QL-induced STAT3 phosphorylation at Ser 727 (Fig. 2C). Interestingly, G␣ 16 QL-induced STAT3 phosphorylation at Tyr 705 was also suppressed by U0126 and the Raf-1 inhibitor, but not by SB202190 and SB203580 (Fig. 2C). Since ERK is a Ser/Thr-specific kinase, these results implied that the ERK pathway might indirectly activate STAT3 by phosphorylation at Tyr 705 .
Roles of Small GTPases on G␣ 16 QL-induced STAT3 Activation-Next, we attempted to identify other intracellular signaling molecules involved in G␣ 16 QL-induced STAT3 activation. Small GTPases have been shown to play critical roles in growth regulation and there is considerable evidence to suggest that GPCR activation can regulate cell growth through the engagement of these monomeric GTPases. In order to determine the role of small GTPases on STAT3 phosphorylation, constitutively activated mutants of Ras (RasG12V), Cdc42 (Cdc42G12V), Rac1 (Rac1G12V), and RhoA (RhoAG14V) were utilized to examine their effects on STAT3 activity. STAT3-dependent luciferase activity was significantly stimulated in cells expressing constitutively active RasG12V or Rac1G12V (Fig. 3A). In contrast, neither Cdc42G12V nor RhoAG14V was able to stimulate STAT3 activity (Fig. 3A).
The functional role of Ras and Rac1 in the regulation of G␣ 16 QL-induced STAT3 activation was further examined in HEK 293 cells. If these small GTPases serve as intermediate signaling molecules for G␣ 16 -mediated STAT3 activation, their dominant negative mutants should inhibit G␣ 16 QL-induced STAT3 activity. Using STAT3 driven luciferase reporter gene assays, we coexpressed the various mutants of Ras and Rac1 with either G␣ 16 or G␣ 16 QL in HEK 293 cells and determined the luciferase activity of the transfectants. In the G␣ 16 -transfectants, the presence of RasG12V significantly elevated STAT3 activity as compared with either the vector control or wild-type Ras (Fig. 3B). The dominant negative mutant of Ras (RasS17N) slightly suppressed the STAT3 activity in G␣ 16expressing cells. In HEK 293 cells expressing G␣ 16 QL and a vector control or together with wild-type Ras, STAT3 activities were significantly higher than the corresponding cells expressing G␣ 16 (Fig. 3B). Expression of constitutively active RasG12V did not further enhance the G␣ 16 QL-induced STAT3 activity, whereas the dominant negative mutant RasS17N almost completely inhibited the G␣ 16 QL-induced STAT3 activity, to a level similar to that seen with G␣ 16 -expressing cells (Fig. 3B). The inhibitory effect of RasS17N on G␣ 16 QL-induced STAT3 phosphorylation was also demonstrated by immunoblotting with anti-phospho-STAT3 antisera. As compared with the vector control or G␣ 16 -transfectants, much higher levels of Tyr 705 and Ser 727 phosphorylation of STAT3 were observed with HEK 293 cells expressing G␣ 16 QL alone, RasG12V alone, RasG12V with G␣ 16 , or G␣ 16 QL together with either wild-type Ras or RasG12V (Fig. 3C). In the presence of RasS17N, the G␣ 16 QLinduced Tyr 705 and Ser 727 phosphorylation of STAT3 were completely suppressed (Fig. 3C). These studies confirmed that activation of Ras leads to STAT3 phosphorylation and that Ras is downstream of G␣ 16 in the STAT3 activation pathway. Similar results were obtained with the use of Rac1G12V and Rac1T17N (a dominant negative mutant of Rac1), although some differences were observed (Fig. 3, B and C). Unlike RasG12V, constitutively active Rac1G12V was unable to induce STAT3 phosphorylation at Ser 727 (Fig. 3C). Nevertheless, Rac1T17N was able to attenuate G␣ 16 QL-induced phosphorylation as well as the activation of STAT3 (Fig. 3C).
Roles of Protein Kinases on G␣ 16 QL-induced STAT3 Activation-The participation of Ras, Rac1, and ERK in G␣ 16 QLinduced STAT3 activation raises the possibility that this signaling pathway may require other molecules directly or indirectly associated with these molecules. To gain further insight into the mechanism of G␣ 16 QL-induced STAT3 activation, the effects of a panel of inhibitors for various kinases and enzymes on STAT3 phosphorylation and activation were investigated. As a member of the G q family, G␣ 16 stimulates phospholipase C␤ (PLC␤) and triggers the mobilization of intracellular Ca 2ϩ and the subsequent activation of protein kinase C (PKC) and calmodulin-dependent protein kinase II (CaMKII). Given that this is a primary pathway for G␣ 16  The immunoblots shown represent one of three sets of immunoblots; two other sets yielded similar results. *, the level of STAT3 phosphorylation was significantly higher than pcDNA1 control (Dunnett t test, p Ͻ 0.05). ϩ, STAT3 phosphorylation at Tyr 705 was significantly higher than that obtained with pcDNA1 control (Dunnett t test, p Ͻ 0.05). #, STAT3 phosphorylation at Tyr 705 was significantly lower than that obtained with pcDNA1 control (Dunnett t test, p Ͻ 0.05). sponses, we examined the possible requirement of PLC␤, PKC, and CaMKII in G␣ 16 QL-induced STAT3 activation. Inhibition of PLC␤ by 10 M U73122 abolished the ability of G␣ 16 QL to induce STAT3-dependent luciferase activity as well as the induction of STAT3 phosphorylation at both Tyr 705 and Ser 727 (Fig. 4A). Similar results were obtained with the CaMKIIspecific inhibitor, KN62 (10 M ; Fig. 4A), and two inhibitors of PKC, 200 nM staurosporin and 100 nM calphostin C (Fig. 4B). Given that some of these inhibitors can exhibit non-selective actions, we further employed kinase-deficient mutants to determine the role of PKC in G␣ 16 QL-induced STAT3 activation. The ␣ and ⑀ isoforms of PKC were selected as representatives of Ca 2ϩ -dependent (26) and -independent (27) members. In HEK 293 cells expressing the kinase-deficient mutant of either PKC␣ (PKC␣-KR) or ⑀ (PKC⑀-KR), G␣ 16 QL-induced STAT3 activity was suppressed to control levels, and phosphorylation levels of Tyr 705 and Ser 727 were not significantly different than those of the G␣ 16 and vector controls (Fig. 4C). Collectively, these results suggest that activation of STAT3 by G␣ 16 QL probably requires signaling via the PLC␤ pathway.
In cytokine receptor signaling, stimulation of STATs is mediated by the phosphorylation and activation of JAKs (7). Interestingly, a number of GPCRs have now been shown to activate JAK2 (14 -16) and JAK3 (28) that subsequently associate with STAT proteins. Hence, we examined the roles of JAK2 and JAK3 in G␣ 16 QL-mediated STAT3 activation using selective inhibitors against these tyrosine kinases. As shown in Fig. 4A, G␣ 16 QL-induced STAT3 activity and its phosphorylation at Tyr 705 and Ser 727 were significantly inhibited by 100 M AG490, a specific inhibitor of JAK2, and by 100 g/ml of a JAK3 inhibitor (WHI-P131). These results suggest that JAK2/3 serve as linkages from G␣ 16 QL to STAT3. This possibility is appealing in view of the fact that both JAK2 and JAK3 can modulate the activity of ERK (29), which appears to participate in G␣ 16 QL-induced STAT3 activation (Fig. 3).
Role of Src Family Tyrosine Kinase on G␣ 16 QL-induced STAT3 Activation-Receptor tyrosine kinases such as the epidermal growth factor (EGF) receptor have long been implicated in Ras-dependent mitogenesis and activate STAT3 (30). Moreover, it has recently been demonstrated that GPCRs can induce cellular responses by transactivating the EGF receptor. We therefore studied the effect of an inhibitor of the EGF receptor, AG1478, on G␣ 16 QL-induced STAT3 activation. The G␣ 16 QLinduced STAT3-dependent luciferase activity was resistant to 50 nM AG1478 (data not shown), indicating that transactivation of the EGF receptor was not necessary for G␣ 16 QL-induced STAT3 signaling in HEK 293 cells. Under similar experimental conditions, AG1478 effectively blocked the phosphorylation and activation of the EGF receptor (results not shown).
Non-receptor tyrosine kinases such as Src have been shown to play a role in cellular proliferation, migration and differentiation. It has been demonstrated that transformation of mammalian fibroblasts by v-Src specifically induced constitutive activation of STAT3 (31), and Src participates in GPCR-induced MAPK activation (32). It was therefore of particular interest to study the action of Src tyrosine kinase on G␣ 16 QLinduced STAT3 activation. To examine the possible involvement of Src, we employed a Src tyrosine kinase inhibitor, PP2, in immunoblotting and reporter gene assays. PP2 dose-dependently inhibited G␣ 16 QL-induced STAT3 phosphorylation at both Tyr 705 and Ser 727 (Fig. 5A). At or above 25 M, PP2 completely suppressed G␣ 16 QL-induced phosphorylation of STAT3-Tyr 705 and STAT3-Ser 727 . G␣ 16 QL-induced STAT3-dependent luciferase activity was also significantly suppressed upon treatment with 25 M PP2, although the inhibition was incomplete (Fig. 5B). Similar results were obtained with an-other Src kinase inhibitor, radicicol (10 M; data not shown). It may be possible that the observed PP2-induced reduction of STAT3 activation was a direct outcome of c-Src inhibition as activation of STAT3 by c-Src has been reported (8). Alternatively, the observed inhibitory effect of PP2 on G␣ 16 QL-induced STAT3 activation may be mediated by a blockade of the MAPKs pathway. The involvement of c-Src in G␣ 16 QL-induced STAT3 activation was further investigated by using a dominant negative mutant of c- Src (c-SrcDN). Coexpression of c-SrcDN significantly attenuated the ability of G␣ 16 QL to induce STAT3 phosphorylation and STAT3-dependent luciferase activity (Fig. 5C), confirming the participation of c-Src in this signaling pathway.
Given that c-Src has been shown to be directly activated by G␣ s and G␣ i , but not by G␣ q , G␣ 12 , or G␤␥ (33), we asked if G␣ 16 can associate with c-Src directly. We assessed the physical association between endogenous c-Src and G␣ 16 QL by coimmunoprecipitation from HEK 293 transfectants. We treated the cell lysates with a chemical cross-linker (Dithiobis[succinimidylpropionate], DSP) prior to immunoprecipitation. There was no evidence of close physical association between G␣ 16 QL and c-Src (data not shown), suggesting that, like G␣ q (33), G␣ 16 cannot associate with c-Src.
Phosphatidylinositol 3-kinase (PI3K) is an important regulatory protein involved in diverse signaling pathways for the control of cellular functions. Non-receptor tyrosine kinases including Src have been shown to activate PI3K. In addition, small GTPases such as Ras, Rac1, and Cdc42 are also known to interact with PI3K, and PI3K has recently been shown to mediate wortmannin-sensitive activation of MAPKs by GPCRs (34). Thus, PI3K appears to be a prime candidate for signal integration in the activation of STAT3 by G␣ 16 QL. In the present study, the role of PI3K on G␣ 16 QL-induced STAT3 activation was examined using the PI3K inhibitors wortmannin and LY294002. G␣ 16 QL-induced STAT3 phosphorylation and activation were unaffected by 100 nM wortmannin or 10 M LY294002 treatment (Fig. 5D), indicating that PI3K was not an essential regulator for STAT3 activation by G␣ 16 QL. The lack of involvement of PI3K in G␣ 16 QL-induced STAT3 activation was further confirmed with the use of a dominant negative mutant of PI3K␥ (PI3K␥DN). Indeed, coexpression of PI3K␥DN with G␣ 16 QL had no effect on the ability of G␣ 16 QL to induce STAT3-dependent luciferase activity (Fig. 5E).
C5a Induces STAT3 Phosphorylation in a PTX-insensitive Manner in HEL Cells-Lastly, we examined whether STAT3 phosphorylation can be stimulated by G␣ 16 -coupled receptors in human erythroleukemia (HEL) cells which are known to express G␣ 16 (35). The complement C5a receptor (C5aR) is capable of activating G␣ 16 (36) and is expressed in hematopoietic cells (37). We therefore treated HEL cells with 100 nM C5a and determined the phosphorylation status of STAT3 by using anti-phosphospecific antibodies. As illustrated in Fig. 6A, C5a significantly stimulated the phosphorylation of STAT3 at both Tyr 705 and Ser 727 residues as compared with the untreated cells. Since the C5aR can utilize both PTX-sensitive and -insensitive G proteins for signal transduction, the possible involvement of G i proteins in mediating the C5a-induced STAT3 phosphorylation was eliminated by pretreating the HEL cells with PTX (100 ng/ml, overnight). C5a remained fully capable of inducing STAT3 Tyr 705 and Ser 727 phosphorylations in PTXtreated cells (Fig. 6A), suggesting that these C5a-induced responses were mediated by PTX-insensitive G proteins such as G␣ 16 . If G␣ 16 was indeed responsible for mediating the C5ainduced STAT3 activation in HEL cells, then it may require the same signaling intermediates as in HEK 293 cells. Indeed, C5a-induced STAT3 Tyr 705 and Ser 727 phosphorylations were significantly inhibited by AG490, Raf-1 inhibitor, U0126, and KN62 in PTX-treated HEL cells (Fig. 6B). These findings illustrated that C5a-induced STAT3 phosphorylations in HEL cells required JAK2, Raf-1, ERK, and CaMKII. DISCUSSION G proteins are major players in numerous biological processes. In addition to their classical role in second messenger activation, they have recently been implicated in the regulation of gene expression. The hematopoietic-specific G␣ 16 can regulate cell differentiation and apoptosis through modulation of MAPKs (3) and transcription factors such as NF-B (38). Since expression of G␣ 16 is regulated during human myeloid differentiation as well as following T-cell activation (39), it represents a prime candidate for the regulation of STAT pathways that are often associated with cytokine signaling. We have recently demonstrated that activation of G␣ 16 -coupled formyl peptide and opioid receptor-like receptors leads to the phosphorylation and activation of STAT3 (23). The present study confirms that G␣ 16 , like G␣ o and G␣ i2 (4), is indeed capable of activating STAT3 through a complex mechanism involving multiple intermediates.
Expression of the constitutively active G␣ 16 QL in HEK 293 cells reproducibly resulted in increased phosphorylations of STAT3 at Tyr 705 and Ser 727 as well as an induction of STAT3dependent luciferase activity. G␣ 16 QL-induced Tyr 705 phosphorylation appeared to be more pronounced than that of Ser 727 because the latter site exhibited higher basal phosphorylation. Phosphorylation of Tyr 705 is required for cytokine-induced STAT3 dimerization, nuclear translocation, and DNA binding, while full transcriptional activity of the homodimer is manifested only when Ser 727 in the transactivation domain is also phosphorylated (40). Since G␣ 16 QL induced STAT3 phosphorylation at both sites, STAT3-dependent transcriptional activity was easily detected by luciferase reporter gene assays. As a signal transducing GTPase, G␣ 16 QL must rely on other tyro-sine and serine kinases to activate STAT3. By using a combination of selective inhibitors and constitutively active and dominant negative mutants, a number of signaling molecules were found to participate in G␣ 16 QL-induced activation of STAT3 (Fig. 7). They included an effector (PLC␤), small GTPases (Ras and Rac1), non-receptor tyrosine kinases (c-Src and JAK2/3), serine kinases (Raf-1, PKC⑀, PKC␣, and CaMKII), and MAPK (ERK). Most of these molecules can be aligned into the Ras/ Raf/MEK/ERK and c-Src/JAK pathways that directly or indirectly phosphorylate STAT3 at Ser 727 and Tyr 705 , respectively.
Although there is increasing evidence to support the participation of STAT3 in G protein-regulated pathways, detailed mapping of the molecular mechanisms involved in such pathways has yet to be completed. Nevertheless, ample evidence is available to support the modulatory role of MAPKs in STAT3 regulation (41). In the present study, we have investigated the regulatory roles of all three members of MAPKs on G␣ 16 QLinduced STAT3 transcription. Our data suggested that only the ERK cascade participated in G␣ 16 QL-induced activation of STAT3, despite the fact that G␣ 16 QL also stimulated JNK and p38 MAPK. Participation of the ERK cascade was supported by several observations. Firstly, ERK was activated by G␣ 16 QL and overexpression of ERK1, but not JNK1 or p38 MAPK, in HEK 293 cells was associated with increased STAT3-dependent luciferase activity. Modulation by ERK of the transcriptional activity of STAT3 via Ser 727 phosphorylation has indeed been reported (10). Secondly, inhibition of MEK1/2 by U0126 led to the suppression of G␣ 16 QL-induced phosphorylation and activation of STAT3. Moreover, the inhibitory effect of a Raf-1 inhibitor on G␣ 16 QL-induced STAT3 transcription highlighted the requirement of Raf-1 (acting as an MEKK in the ERK cascade) in STAT3 activation. Lastly, the involvement of Ras was demonstrated with the use of wild-type and mutant Ras; RasG12V alone stimulated STAT3 activities while RasS17N suppressed the G␣ 16 QL-induced phosphorylation and activation of STAT3. Indeed, it has recently been reported that G␣ 16 QL can activate Ras via direct interaction with a novel adaptor protein known as tetratricopeptide repeat 1 (35). Constitutively active Ras protein has long been associated with pathogenesis of human cancer. The role of this small GTPase in signal transduction has been extensively studied and its crucial role in the initiation of the ERK cascade is well established. The stimulatory effect of RasG12V and inhibitory action of RasS17N on STAT3 transcriptional activity further revealed and strengthened the important connection between MAPK pathway and STAT3 activation. Unlike G␣ o -induced STAT3 activation in NIH-3T3 fibroblasts (4), the classical Ras/Raf/ MEK/ERK pathway commonly employed by mitogens thus appeared to play an essential role in G␣ 16 QL-induced STAT3 activation in HEK 293 cells. Interestingly, although neither Ras nor Raf-1 acts as a tyrosine kinase, inhibition of the activity of either Ras or Raf-1 significantly attenuated G␣ 16 QLinduced phosphorylation of STAT3 at Tyr 705 . These observations reiterated the possible existence of complex regulatory mechanisms involved in MAPK-mediated regulation of STAT proteins.
Apart from Ras, the Rho family of GTPases has also been shown to regulate STAT3 activity. Activation of STAT3 by Rac1 has been reported (11) and this pathway is actually employed by GPCRs (42). Recently, Rac1 has been shown to induce STAT3 phosphorylation directly (43). Complementary to the existing literature, we have also demonstrated an increase in STAT3 phosphorylation and activation induced by constitutively active Rac1. The dominant negative mutant of Rac1, on the other hand, prevented G␣ 16 QL from activating STAT3. However, not all small GTPases are involved in STAT3 regulation. RhoA and Cdc42 did not affect STAT3 activation in the present study. In view of the effects of various GTPases on STAT3 activity, it is likely that small GTPases may play a key role in network signaling. By acting as signal activators, small GTPases may provide a significant linkage between different signaling pathways. Ras and Rac1 might eventually converge at the level of ERK since Rac1G12V has been shown to activate ERK in Rat-2 fibroblasts (44). It is believed that intimate interactions between different signaling pathways contribute to the appropriate control of intracellular signaling (45).
One of the mechanisms by which heterotrimeric G proteins regulate the activities of monomeric GTPases is via second messengers. As a member of the G q family, G␣ 16 QL constitutively stimulates the activity of PLC␤ (18), subsequently leading to the activation of downstream effectors such as PKC and CaMKII. Given that G␣ 16 QL-induced STAT3 activation was effectively inhibited by U73122, KN62, staurosporin and calphostin C, PLC␤ and its downstream effectors might constitute an important signaling pathway in mediating the actions of G␣ 16 QL. Linkages to STAT3 from the PLC␤ pathway can be traced back to ERK. Group I metabotropic glutamate receptors have been shown to activate ERK via CaMKII in striatal neurons (46), and CaMKII-dependent activation of ERK in vascular smooth muscle requires the participation of tyrosine kinases such as Pyk2 and Src (47). As for PKC-dependent activation of ERK, gonadotrophin-releasing hormone-induced phosphorylation of ERK in hypothalamic neurons also requires Pyk2 and Src (48). Gonadotrophin-releasing hormone stimulation caused the translocation of PKC␣ and PKC⑀ to the cell membrane and enhanced the association of Src with PKC␣ and PKC⑀. The same two isoforms of PKC appeared to play a role in G␣ 16 QL-induced STAT3 activation because their dominant negative mutants were effective in suppressing the phosphorylation and activation of STAT3. It should also be noted that PKC and CaMKII can also activate ERK via Raf-1 (49) and Ras (50), respectively. Indeed, both PKC␣ (51) and PKC⑀ (52) have been shown to activate Raf-1, leading to the stimulation of ERK. Irrespective of the loci for signal integration, activation of STAT3 by G␣ 16 QL apparently involved a number of intermediates along the PLC␤ pathway.
There is considerable evidence to support the involvement of cross talk between multiple signaling pathways and certain protein tyrosine kinases are thought to play important cooperative roles in mitogenic STAT3 signaling (53). Non-receptor tyrosine kinases such as JAKs have been shown to participate in signaling from a range of cell-surface receptors such as those for cytokines and growth factors. Activation of these receptors results in the recruitment and activation of JAKs, leading to subsequent binding, phosphorylation and activation of STAT3 (7). In addition to JAKs, the Src family of non-receptor kinases has been implicated in the phosphorylation and activation of STAT proteins (8). The inhibitory effects of PP2, AG490 and JAK3 inhibitor on STAT3 activation observed in the present study verified the important roles of c-Src and JAK2/3 in G␣ 16 QL-induced STAT3 activation. Involvement of JAKs in c-Src-dependent STAT3 activation has previously been shown in PDGF-induced signaling (53). It is possible that G␣ 16 QLinduced STAT3 activation may involve cooperative association between c-Src and JAKs. The ability of G␣ 16  JAKs provides a means to phosphorylate STAT3 at Tyr 705 , while Ser 727 can be phosphorylated via the Ras/Raf/MEK/ERK axis. Moreover, the involvement of JAK2/3 in mediating the G␣ 16 QL signals to STAT3 is in good agreement with their ability to activate Ras and ERK (54).
A requirement for c-Src in G␣ 16 QL-induced STAT3 activation can be predicted from the Src-dependences of G␣ i2 -and G␣ o -induced stimulation of STAT3 (4). As a ubiquitously expressed non-receptor tyrosine kinase, c-Src is well positioned to serve as a key signaling molecule in a variety of pathways. The inhibitory effects of the selective inhibitor and the dominant negative c-Src mutant demonstrate the involvement of c-Src in G␣ 16 QL-induced STAT3 activation. One of the most obvious consequences of the regulation of c-Src by G␣ 16 QL is the provision of a direct linkage to the JAK/STAT pathway. In addition, direct activation of STAT3 by v-Src has been demonstrated (8,31). Yet, v-Src can also activate Ras and PI3K and thus regulate STAT3 via these intermediates. In G␣ 16 QL-induced STAT3 activation, c-Src did not appear to require PI3K because STAT3 activity was unaffected by wortmannin, LY294002, and PI3K␥DN. As discussed earlier, the Ras/Raf/ MEK/ERK cascade probably lies downstream of c-Src in G␣ 16 QL-induced STAT3 activation and the ability of c-Src to mediate GPCR-induced activation of ERK has been well established (55). Additionally, binding of c-Src to activated G␣ subunits has been demonstrated for G␣ s and G␣ i , but not for G␣ q , G␣ 12 , or G␤␥ (33). The lack of effect of constitutively active mutants of G␣ q and G␣ 12 on STAT3 activity in NIH-3T3 cells (4) is in line with their inability to activate c-Src and suggests that c-Src may serve as a crucial signaling molecule in G␣mediated activation of STAT3. Despite the fact that G␣ 16 QLinduced STAT3 activation requires the participation of c-Src, there is apparently no physical association between G␣ 16 QL and c-Src, supporting the view that G␣ q family members are incapable of directly activating c-Src. In this regard, G␣ 16 QL must employ a mechanism that is different from that of G␣ i2 QL and G␣ o QL to stimulate STAT3 activity. It remains to be determined if adaptor proteins can provide a linkage between c-Src and G␣ 16 QL.
As a member of a family of latent cytoplasmic transcription factors, STAT3 has long been implicated in the normal processes of cell growth and development. More recently, evidence for a role of STAT3 in oncogenesis has become increasingly apparent. Constitutive activation of STAT3 in v-Src transformed fibroblasts (8), the imperative requirement of STAT3 in v-Src induced transformation (31), and cellular transformation induced by constitutively activated STAT3 (56), have provided strong evidence to support the tumorigenic potential of STAT3. In NIH-3T3 cells, G␣ i2 -and G␣ o -induced neoplastic transformation is mediated via Src and STAT3 (4,16). In contrast to G␣ i2 and G␣ o , persistent activation of G␣ 16 leads to inhibition of cell growth in Swiss 3T3 fibroblasts (18) and to cell differentiation in PC12 cells (3), probably due to prolonged activation of JNK. It remains to be determined if activation of STAT3 by G␣ 16 can indeed lead to neoplastic transformation in specific cell types. It is noteworthy that the mechanisms by which G␣ i2/o and G␣ 16 activate STAT3 differ significantly in their dependences on ERK, although both mechanisms require c-Src. Importantly, the ability of G␣ 16 to activate STAT3 (and perhaps STAT1) offers GPCRs the opportunity to modulate cytokine signaling in hematopoietic cells. In this regard, the chemokine CXCR4 receptor has been shown to stimulate the JAK/ STAT pathway independently of G i proteins in human T cells (28). Since G␣ 16 is expressed in T cells (39), this response may well be mediated by G␣ 16 . In HEL cells that express both G␣ 16 and the C5aR (35,37), we demonstrated that the C5a-induced STAT3 phosphorylation was PTX-insensitive (Fig. 6). Since the C5aR is capable of stimulating both G i/o and G␣ 16 , the observed PTX-insensitivity verified that C5a-induced STAT3 phosphorylations in HEL cells are G i/o -independent and may well be regulated by G␣ 16 . Although we did not unequivocally demonstrate the involvement of G␣ 16 in mediating the C5a response, the sensitivity of the pathway to various inhibitors suggested that it is similar, if not identical, to the one employed by G␣ 16 QL in HEK 293 cells. Both pathways require the participation of JAK2, Raf-1, MEK1/2, and CamKII. Nevertheless, additional studies are needed to confirm the involvement of G␣ 16 in GPCR-induced STAT3 activation in hematopoietic cells. Such studies would necessitate the use of G␣ 16 -specific siRNA or antibodies. The demonstrated ability of G␣ 16 QL to stimulate STAT3 activity in HEK 293 cells provides a mechanism by which C5a may elicit its potent proinflammatory actions.
Based on our observations, it is unlikely that a single signaling pathway is entirely responsible for G␣ 16 QL-induced STAT3 activation. Involvement of the c-Src/JAK and ERK cascades in STAT3 activation was evidently significant, and even perhaps mutually complementary (Fig. 7). It has been speculated that selective engagement of distinct signaling pathways is required for the elicitation of specific biological responses (15). The present study only provides a rudimentary road map for GPCRs to regulate STAT3 activity via G␣ 16 , a number of issues remain to be resolved. Among these is the question of why JNK and p38 MAPK are not essential for G␣ 16 QL-induced STAT3 activation despite their demonstrated stimulatory activity on STAT3 (11). It should also be noted that the use of G␣ 16 QL in this study eliminates any potential contribution by the G␤␥ complex. Given that G␤␥ can regulate numerous effectors including PLC␤ and the MAPKs, activation of STAT3 by G␣ 16 -coupled receptors may not utilize the same set of signaling molecules. The incessant discovery of unique signaling pathways and identification of crucial signaling molecules will greatly enhance our current understanding of intracellular mechanisms involved in heterotrimeric G protein signaling.