Gi Proteins Use a Novel βγ- and Ras-independent Pathway to Activate Extracellular Signal-regulated Kinase and Mobilize AP-1 Transcription Factors in Jurkat T Lymphocytes

Receptors coupled to pertussis toxin (PTX)-sensitive Gi proteins regulate T lymphocyte cytokine secretion, proliferation, and chemotaxis, yet little is known about the molecular mechanisms of Gi protein signaling in mammalian lymphocytes. Using the Jurkat T lymphocyte cell line, we found that a stably expressed Gi protein-coupled receptor (the δ-opioid receptor (DOR1)) stimulates MEK-1 and extracellular signal-regulated kinases 1 and 2 (ERK1 and ERK2) and transcriptional activity by an ERK target, Elk-1, via a mechanism requiring a PTX-sensitive Gi protein. Levels of β-adrenergic receptor kinase-1 C-terminal fragment that inhibited signaling by Giprotein βγ subunits in these cells had no effect on DOR1 stimulation of either MEK-1- or Elk-1-dependent transcription, indicating that this pathway is independent of βγ. Analysis of this βγ-independent pathway indicates a role for a herbimycin A-sensitive tyrosine kinase. Unlike βγ-mediated pathways, the βγ-independent pathway was insensitive to RasN17, inhibitors of phosphatidylinositol 3-kinase (PI 3-kinase), and constitutive PI 3-kinase activity. The βγ-independent pathway regulates downstream events, since blocking it abrogated both Elk-1-dependent transcription and mobilization of the mitogenic transcription factor, AP-1, in response to DOR1 signaling. These results characterize a novel, Ras- and PI 3kinase-independent pathway for ERK activation by Gi protein signaling that is distinct from ERK activation by βγ and may therefore be mediated by the αi subunit.

G i protein signaling in many cell types stimulates the activity of the extracellular signal-regulated kinases (ERKs), 1 ERK1 and ERK2 (Refs. 1-16; reviewed in Ref. 17). ERK and other mitogen-activated protein kinases phosphorylate transcription factors, thereby mediating transcriptional regulation in response to signaling by cell surface receptors (18 -22). Like other G proteins, G i proteins consist of an ␣ subunit and a dimeric ␤␥ subunit and mediate activation of intracellular signaling pathways in response to the ligation of receptors with seven transmembrane-spanning domains. When stimulated by a ligandbound receptor, the ␣ subunit binds GTP and dissociates from ␤␥. This event allows both ␣-GTP and free ␤␥ to directly regulate the activities of downstream effector molecules. G i proteins contain one of the closely related ␣i subunits (␣i1, ␣i2, ␣i3) in addition to ␤ and ␥ subunits shared by other types of G proteins. Pertussis toxin (PTX) covalently modifies G i protein ␣i subunits, preventing G i protein activation by ligated receptors. Besides G o and G t , only G i proteins are inactivated by PTX.
The ␣ subunits of PTX-sensitive G proteins also have the potential to activate ERK. The gip2 oncogenes encode constitutively active ␣i2 subunits with either an R179C or a Q205L point mutation that inhibits GTP hydrolysis (24). gip2 expression constitutively stimulates ERK (25) and mediates the oncogenic transformation of Rat1a fibroblasts (24,26). Unlike ␤␥ stimulation of ERK (1)(2)(3)22), gip2 stimulates ERK via a mechanism that is independent of Ras (25), a result suggesting that ␣i and ␤␥ activate ERK via distinct pathways. Moreover, the m1-muscarinic and platelet-activating factor receptors use the closely related PTX-sensitive ␣o subunit to activate ERK via a Ras-independent pathway in the Chinese hamster ovary epithelial cell line (10). An interesting feature of this pathway is its cell type specificity; in COS-7 fibroblasts, these same receptors use ␤␥ and Ras to stimulate ERK (10). This resembles the cell-specific actions of gip2, which transforms Rat-1a cells but not NIH 3T3 or Swiss 3T3 fibroblasts (24). Recent evidence suggests that ␣i and ␣o activate ERK in Rat-1a and COS-7 cells via transactivation of the epidermal growth factor receptor tyrosine kinase (7,13). Although Src family tyrosine kinases have been implicated in this process (13), additional mechanistic details of this pathway are obscure.
Lymphocytes express abundant heterotrimeric G i proteins (27) that direct lymphocyte migration (28 -31), proliferation (32,33), and cytokine secretion (34,35). Although recent work has uncovered novel mechanisms for G i protein stimulation of ERK activity in fibroblast and epithelial cell types (above), little is known about the molecular mechanisms of ERK activation by G i protein signaling in mammalian lymphocytes. We recently reported that signaling by a heterologously expressed G i protein-coupled ␦-opioid receptor (DOR1) uses a PTX-sensitive mechanism to stimulate the accumulation of AP-1 transcription factor complexes in the Jurkat human T lymphocyte cell line (35). Here, we demonstrate that DOR1 signaling activates ERK via a novel, Ras-independent, G i protein-mediated pathway. We further present evidence that this pathway is independent of ␤␥, characterize this pathway as distinct from previously characterized ␤␥-mediated ERK activation pathways, and demonstrate that this pathway leads to PTX-sensitive AP-1 mobilization in these cells.

EXPERIMENTAL PROCEDURES
Cells-DOR1/Ju.1 cells were derived by fluorescence-activated cell sorting of the Ju.1 Jurkat T cell subline that was stably transfected with DOR1, as described previously (36). All cells were maintained in Medium B (RPMI 1640 supplemented with 5% fetal calf serum, 5% calf serum, 10 mM HEPES, pH 7.4, 2 mM L-glutamine, and 2 M 2-mercaptoethanol). DOR1 expression by the DOR1/Ju.1 cell line was confirmed by assays every few months as described (35).
Transfections and Assays for Elk-1-dependent Transcriptional Activity-Elk-1 dependent transcriptional activity was assayed using the PathDetect Elk trans-Reporting System (Stratagene, La Jolla, CA). Cells (10 7 ) were mixed with 5 g of pFR-Luc and 0.5 g of pFA-Elk with or without transient expression constructs described below plus irrelevant plasmid DNA (pcDNAIII; Invitrogen, San Diego, CA), electroporated using a BTX Electro-square porator model T820 (BTX Inc., San Diego, CA), divided into 4 -6 portions, and cultured in 2 ml of Medium B or Medium C in wells of a 24-well plate for 18 -24 h at 37°C. Results were similar following culture in Medium B or C, which contain 10% and 0.5% serum, respectively. Stimuli were added to some wells for an additional 5-6 h, and then the cells were harvested and assayed for luciferase using the Luciferase Assay System (Promega, Madison, WI) and a model LB 9501/16 lumat (Berthold Systems, Aliquippa, PA). DOR1-stimulated Elk1/GAL4 activity was calculated relative to that of unstimulated cells from the same transfection. To compare different transfections done the same day, the cells were additionally transfected with 20 ng of pRL-TK (Promega, Madison, WI), which encodes a Renilla luciferase gene downstream of a minimal HSV-TK promoter. Following stimulation of transfected cells, both Renilla luciferase and luciferase were measured using the Dual Luciferase Assay System (Promega), and the Renilla values were used for normalization.
In Vitro Kinase Assays of Overexpressed, Myc-tagged MEK-1-Cells (10 7 ) were transfected with 5 g of an expression vector encoding Myc-tagged MEK-1 (Myc-MEK-1) with or without other expression plasmids and plated in Medium B. 24 h later, Myc-MEK-1 was immunoprecipitated using an anti-Myc mAb (Babco, Berkley, CA), and an in vitro kinase assay was performed as described (43) in the presence of [␥-32 P]ATP, using as a substrate 1 g/test of kinase-inactive ERK2 GST fusion protein (GST-ERK2-KD; prepared as described (44)). Results were visualized by SDS-PAGE, Western transfer to Immobilon P membrane, and autoradiography, and/or PhosphorImager analysis using a Storm model 840 (Molecular Dynamics, Inc., Sunnyvale, CA). Quantitation was performed by PhosphorImager analysis. Total immunoprecipitated Myc-MEK-1 was visualized by immunoblotting with anti-MEK-1 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA).
Infection with Recombinant Vaccinia Viruses-To make the recombinant Vaccinia viruses, the coding sequence of human CD56 or human Ha-RasN17 was blunt end-cloned into the SmaI site of the Vaccinia virus expression vector, pSC11. The resulting vectors were introduced into the WR strain of Vaccinia via homologous recombination, and high titer viral stocks were prepared as described (45). Stocks of the WR (wild-type/parental) strain Vaccinia virus were a generous gift of Dr. Paul Leibson (Mayo Clinic, Rochester, MN). Mutant Ras expression by the RasN17 recombinant virus was confirmed by reverse transcriptasepolymerase chain reaction and DNA sequencing of the viral stock (data not shown) and by anti-Ras immunoblotting of infected cells (Fig. 5, B and C). Semipurified recombinant CD56, RasN17, or WR (control) viruses were used for infection. For each experiment, 4 ϫ 10 7 cells were infected in 8.0 ml of RPMI 1640 for 1 h at 37°C at a multiplicity of infection of 20:1. One volume of RPMI 1640 supplemented with 1% fetal calf serum was then added, and the cells were incubated for an additional 1 h at 37°C. Infected cells were washed twice with RPMI 1640 supplemented with 0.5% fetal calf serum and either immediately incubated with anti-CD56 mAb coupled to phycoerythrin (Becton-Dickenson, San Jose, CA) and analyzed by flow cytometry or divided into six equal portions and used for endogenous MEK-1 and/or ERK2 kinase assays as described below.
Kinase Assays of Endogenous ERK2 and MEK-1-Cells (2 ϫ 10 6 /test) were stimulated as indicated at 37°C in medium consisting of RPMI 1640 supplemented with 1% bovine serum albumin and lysed as described by Kohn et al. (43). Endogenous MEK-1 and/or ERK2 was immunoprecipitated using specific antisera (Santa Cruz Biotechnology) and in vitro kinase assays performed as for the Myc-MEK-1 assay above, using as substrate either 1 g/test of kinase-inactive ERK2-GST fusion protein (for MEK-1 assay) or 5 g/test of myelin basic protein (MBP) (for ERK2 assay; Upstate Biotechnology, Inc., Lake Placid, NY). Immunoblotting was used to quantitate the amounts of ERK2 or MEK-1 immunoprecipitated for each test (anti-ERK2 and anti-MEK-1 antisera from Santa Cruz Biotechnology). Where indicated, 1 ⁄20 of the total lysate was reserved and analyzed by immunoblotting for RasN17 expression using anti-Ha-Ras antisera (Santa Cruz Biotechnology).
Akt Kinase Assay-Cells (2 ϫ 10 7 ) were transfected with 5 g of an expression vector encoding AU1-tagged Akt (pEF-BOS-AU1-AKT) with or without 10 g of pEF-BOS-FLAG-iSH2-CAAX and plated in Medium B. 24 h later, Akt was immunoprecipitated using an anti-AU1 mAb (Babco, Berkley, CA), and an in vitro kinase assay was performed using myelin basic protein as a substrate (43). Results were visualized by SDS-PAGE, transfer to Immobilon P membrane, and autoradiography. Immunoprecipitated Akt was visualized by immunoblotting with sheep anti-Akt antiserum and horseradish peroxidase-conjugated anti-sheep IgG (both antisera from Upstate Biotechnology).
RNA Blot-Cells were pretreated and stimulated as indicated in Medium B at a density of 5 ϫ 10 5 to 8 ϫ 10 5 cells/ml, and total RNA was prepared using Trizol reagent according to the manufacturer's instructions (Life Technologies, Inc.). RNA (10 g/lane) was fractionated by electrophoresis in a 1% agarose-formaldehyde gel, transferred to Hybond-N membrane (Amersham Pharmacia Biotech), and hybridized with 32 P-labeled c-Fos and glyceraldehyde-3-phosphate dehydrogenase cDNA probes using standard procedures.
Detection of AP-1 DNA-binding Complexes-Cells were stimulated for 4 h in Medium B at a density of 5 ϫ 10 5 to 8 ϫ 10 5 cells/ml. Nuclear extracts were prepared (35), and AP-1 DNA-binding complexes were assayed by electromobility shift assay as described (37). The doublestranded DNA probe was 5Ј-AAATCCAATGAGTCAGCGCGGAT-3Ј, which contains a high affinity AP-1 binding site (underlined). (10 7 ) in Medium B were transfected with 10 g of irrelevant plasmid DNA (pcDNAIII; Invitrogen) and 20 g of the AP-1 reporter plasmid, pGL-AP1-Luc (generously provided by E. O'Neill (Merck)), which contains three tandem repeats of a sequence (5Ј-GTGACTCAGCGCG-3Ј) with a high affinity AP-1 binding site (underlined) upstream of a minimal ␥-fibrinogen promoter (46) and a luciferase gene. Following culture in Medium C for 18 -24 h at 37°C, cells were pretreated and stimulated for an additional 5-6 h before reporter activity was assayed by luciferase as above.

RESULTS
The G i Protein-coupled Receptor, DOR1, Activates MEK-1, ERK1, and ERK2 in Jurkat T Cells via a PTX-sensitive Pathway-To address the potential for mitogenic signaling by G i proteins in lymphocytes, we first asked whether receptor-mediated activation of endogenous G i proteins could stimulate ERK activity in a lymphoid cell line. For this purpose, we used DOR1/Ju.1 cells, a Jurkat T cell line stably transfected with an expression vector encoding the G i protein-coupled receptor, DOR1 (35,36). ERK activation was measured in response to deltorphin, a specific DOR1 agonist. Fig. 1A shows that treatment with 10 Ϫ7 M deltorphin stimulated the appearance of active, phosphorylated ERK1 and ERK2 in DOR1/Ju.1 cell lysates. This effect of deltorphin required G i proteins since DOR1, but not PMA, activation of ERK was abrogated by pretreatment with PTX. DOR1 activation of ERK1 and ERK2 most likely requires upstream MEK activation, since ERK activation was sensitive to the specific MEK inhibitor, PD098059 ( Fig. 1A) (47). In fact, DOR1 signaling activates MEK-1, shown by transiently transfecting a construct encoding Myc-tagged MEK-1 and then immunoprecipitating and assaying its kinase activity (Fig. 1B). DOR1 signaling also activates the kinase activity of endogenous MEK-1 (Fig. 5C). Like DOR1 activation of ERK1 and ERK2, DOR1 activation of either overexpressed or endogenous MEK-1 was completely sensitive to pretreatment with PTX (data not shown).
Since the transcription factor, Elk-1, is activated by ERK phosphorylation (20,21), we also tested the ability of DOR1mediated G i protein signaling to mediate Elk-1-dependent transcriptional activity. DOR1/Ju.1 cells were transiently transfected with a plasmid that constitutively expresses a chimeric Elk-1/Gal4 transcription factor (pFA-Elk), together with a plasmid containing the luciferase gene downstream of five Gal4 DNA binding sites (pFR-Luc). When activated by upstream signaling events, ERK phosphorylates the Elk-1 domain of the transfected Elk/Gal4 chimera, which then stimulates luciferase expression from the Gal4-responsive promoter of pFR-Luc.
As shown in Fig. 1C, treatment of transiently transfected DOR1/Ju.1 cells with 10 Ϫ7 M deltorphin increased Elk-1-dependent luciferase activity. Elk-1 may potentially be regulated by several mechanisms; however, the deltorphin-mediated response is sensitive to the specific MEK inhibitor, PD098059. In contrast, PD098059 had little effect on luciferase activity derived from a cotransfected control Renilla reporter gene (Con.). Elk-1-dependent transcription in response to deltorphin was dose-dependent (Fig. 1C, inset) in a manner consistent with its reported K d of 3.3 nM for DOR1 (21). Subsequent experiments used 10 Ϫ7 M deltorphin for stimulation. Fig. 1D shows that the Elk-1-dependent transcription stimulated by deltorphin was completely inhibited by pretreatment with PTX but not by pretreatment with the PTX B-oligomer, which lacks the ADPribosyltransferase subunit that targets G protein ␣i subunits. Fig. 1D, inset, shows that these effects of PTX and PTX-B were observed in multiple experiments. In contrast, PTX had no effect on Elk-1-dependent transcription stimulated by PMA. Moreover, deltorphin had no effect on luciferase expression from the control Renilla reporter gene or on luciferase expression in DOR1/Ju.1 cells cotransfected with pFR-Luc and a plasmid encoding only the Gal4 DNA binding domain (data not shown). Together, these results demonstrate that DOR1 uses a PTX-sensitive G i protein to activate MEK-1, ERK1, and ERK2 and to mediate transcriptional activation by an ERK target, Elk-1, in Jurkat T cells.

Overexpression of G i Protein ␤␥ Subunits Stimulates MEK-1 Activity and Elk-1-dependent Transcription in Jurkat T Cells-
Overexpression of the G protein ␤ 1 and ␥ 2 subunits mimics ␤␥ signaling that follows the ligation of some G i protein-coupled receptors (1-6, 8, 10, 12). We therefore examined the ability of ␤␥ to activate Elk-1-dependent transcription in Jurkat T cells. Cells were transiently transfected with pFR-Luc and pFA-Elk, together with expression vectors encoding ␤ 1 and ␥ 2 . Fig. 2A shows typical results. Transient expression of both ␤ 1 and ␥ 2 subunits significantly elevated Elk-1-dependent luciferase activity (solid bars) without affecting expression from the control Renilla reporter gene (open bars, Con.). In contrast, transient expression of either ␤ 1 or ␥ 2 alone had little effect on Elk-1-dependent or control luciferase activity. In multiple experiments, overexpressing both ␤ 1 and ␥ 2 elevated Elk-1-dependent transcriptional activity by 5.3 Ϯ 0.6-fold compared with basal levels (n ϭ 7; p Ͻ 0.01) (Fig. 2B). Elk-1-dependent transcriptional activity that was stimulated by overexpressed ␤␥ was sensitive to the MEK inhibitor, PD098059, but PD098059 had little effect on the transcription of luciferase from the control Renilla reporter gene (Fig. 2, A and B, and data not shown).
To directly examine the regulation of MEK activity by overexpressed ␤␥, we transiently transfected ␤␥ together with a construct encoding Myc-tagged MEK-1 and then immunoprecipitated and assayed the kinase activity of the Myc-tagged MEK-1. Results shown in Fig. 3C show that overexpressing ␤␥ in Jurkat T cells resulted in an approximately 2-fold activation of Myc-MEK-1 kinase activity. These results show that overexpressing G protein ␤␥ subunits in Jurkat T cells increases both MEK-1 activity and Elk-1-dependent transcription, suggesting that ␤␥ is capable of activating ERK in these cells.
We next tested the ability of cotransfected ␤ARKct to inhibit DOR1 stimulation of the Elk-1-dependent transcriptional activity that is downstream of MEK-1 and ERK. While cotransfecting ␤ARKct clearly inhibited the ability of overexpressed ␤␥ subunits to stimulate Elk-1-dependent transcription, ␤ARKct had little effect on the transcriptional activity that resulted from DOR1 ligation by deltorphin (Fig. 4A). Fig. 4B summarizes the results of multiple experiments. ␤ARKct exerted no significant effect on the ability of deltorphin to stimulate Elk-1-dependent transcriptional activity (n ϭ 3; p ϭ 0.96), while in the same experiments ␤ARKct significantly inhibited activity in response to overexpressed ␤␥ (n ϭ 3; p Ͻ 0.02). Together, the results in this section indicate that DOR1 uses a ␤␥-independent mechanism to stimulate MEK-1 kinase activity and transcriptional activation by Elk-1. Since this pathway is also sensitive to PTX (Fig. 1), these results suggest that DOR1 uses ␣i rather than ␤␥ to stimulate MEK-1 and downstream events in Jurkat T cells.
Ras  (Fig. 5A). Using identical infection conditions, we infected Jurkat T cells with recombinant Vaccinia viruses that express either RasN17 or nothing (WR, or wildtype, strain). Following stimulation with deltorphin, the pres- ence of active, phosphorylated ERK1 and ERK2 was assayed by immunoblotting as in Fig. 1A. As shown in Fig. 5B, RasN17 had no detectable effect on the time course of ERK1 or ERK2 phosphorylation in response to deltorphin. RasN17 also had no effect on the ability of deltorphin to activate endogenous MEK-1 or ERK2 activity, as measured by immunoprecipitating these kinases and measuring their kinase activities in vitro (Fig. 5C). Fig. 5D shows that in multiple experiments, we observed no significant inhibition of DOR1-mediated activation of endogenous ERK2 or MEK-1 kinase activity in cells expressing RasN17.
We also tested the effects of RasN17 on DOR1 signaling that leads to Elk-1-dependent transcriptional activity. An expression plasmid encoding RasN17 was transfected into DOR1/Ju.1 cells together with pFR-Luc and pFA-Elk, and the cells were stimulated and assayed for Elk-1-dependent transcriptional activity as in Fig. 1, C and D. Typical results are shown in Fig.  6A. RasN17 inhibited Elk-1-dependent transcription in response to either ␤␥ overexpression or stimulation of the T lymphocyte antigen receptor-CD3 complex (TCR-CD3) with anti-CD3 mAb. In addition, cotransfection of this RasN17-encoding plasmid inhibited (by approximately 80%) transcription from an IL-2 promoter construct in response to TCR-CD3 and CD28 stimulation (n ϭ 3; p ϭ 0.02; data not shown). These results are consistent with previous reports that Ras participates in signaling by ␤␥ (2, 3, 10) and TCR-CD3 (49,50). In contrast to its effects on TCR-CD3 and ␤␥ signaling, RasN17 had little or no effect on Elk-1-dependent transcriptional activity stimulated by DOR1 or PMA (Fig. 6A). Fig. 6B summarizes the results of multiple experiments. RasN17 consistently inhibited ␤␥, but not deltorphin, stimulation of Elk-1-dependent transcription.
To summarize, the results in this section demonstrate that DOR1 uses a RasN17-insensitive mechanism to activate MEK-1-, ERK1-and ERK2-, and Elk-1-dependent transcriptional activity in Jurkat T cells. In contrast, ␤␥ stimulation of MEK-1 and Elk-1-dependent transcription in these cells is sensitive to RasN17. Since DOR1, but not ␤␥, signaling is also insensitive to ␤ARKct, these results suggest that DOR1 and ␤␥ subunits stimulate ERK activation via independent signaling pathways that differ in their requirements for Ras.
Elk-1-dependent Transcription in Response to ␤␥, but Not DOR1, Synergizes with Constitutive PI 3-Kinase Activity-In fibroblast and epithelial cells, ␤␥-dependent stimulation of ERK requires activity of the p110-␥ isoform of PI 3-kinase (8,12,23). We therefore asked whether ␤␥or DOR1-mediated Elk-1-dependent transcription synergizes with constitutive PI 3-kinase activity. The binding of p85 to activated receptor tyrosine kinases normally stimulates the PI 3-kinase activity of associated p110␣/␤ PI 3-kinases by mediating their membrane localization (51, 52). We therefore constructed an expression  4. ␤ARKct has no effect on DOR1 stimulation of Elk-1-dependent transcription. DOR1/Ju.1 Jurkat T cells were transfected with pFR-Luc and pFA-Elk with or without plasmids encoding ␤ 1 , ␥ 2 , or 5 g of ␤ARKct. Deltorphin or ␤␥ stimulation of Elk-1-dependent transcriptional activity was measured as in Figs. 1 and 2. A, results of a representative experiment, with results expressed as a percentage of the PMA response in the same transfection. B, summary of multiple experiments as in A, with deltorphin or ␤␥ stimulation in the absence of ␤ARKct normalized to 100%. Each bar denotes the mean Ϯ S.E. of three independent determinations. *, the mean is significantly different from 100% (p Ͻ 0.02). plasmid encoding iSH2-CAAX, which consists of the p110␣/␤binding iSH2 domain of p85 fused to the membrane targeting and isoprenylation domain of Ha-Ras. Transient expression of iSH2-CAAX enhanced the kinase activity of Akt, a downstream target of PI 3-kinase (53, 54) (Fig. 7A), indicating that like other fusion proteins that mediate membrane localization of p110␣/␤ (51, 52, 55), iSH2-CAAX stimulates PI 3-kinase activity. Consistent with this idea, iSH2-CAAX activation of Akt was abrogated by pretreatment with the PI 3-kinase inhibitor, wortmannin. iSH2-CAAX also mimicks PI 3-kinase signaling when transiently expressed in other cell types (56).
We also tested the effects of the PI 3-kinase inhibitor, wortmannin, on ERK activation by DOR1. As shown in Fig. 7D, wortmannin applied at doses that completely inhibit p110␣/␤ or p110-␥ PI 3-kinase activity (12,57), as well as iSH2-CAAXstimulated Akt activity (Fig. 7A), had little effect on ERK activation by deltorphin. Similarly, Elk-1-dependent transcriptional activity in response to deltorphin was unaffected by pretreatment with wortmannin (data not shown). Since the iSH2-CAAX fusion protein enhanced Elk-1-dependent transcription in response to signaling by ␤␥, but not DOR1, these results provide additional support for the idea that DOR1 and ␤␥ stimulate ERK activity via distinct pathways. In contrast, the results in this section indicate that the DOR1-mediated pathway of ERK activation is independent of PI 3-kinase activity.

MEK-1 Activation Is Essential for DOR1 Signaling That Leads to Increased Levels of c-Fos mRNA and Transcriptionally Active AP-1 Transcription Factors-When
Elk-1 is phosphorylated by ERK, it can act together with serum response factor to mediate increased expression of c-Fos (18 -21). Fig. 8A shows that deltorphin stimulated the accumulation of c-Fos mRNA but not control glyceraldehyde-3-phosphate dehydrogenase mRNA. Furthermore, this increase was abrogated by pretreatment with the MEK inhibitor, PD098059. Since PD098059 also blocks ERK activation and Elk-1-dependent transcription in response to deltorphin (Fig. 1), these results suggest that DOR1 requires MEK-1 and ERK activation to increase c-Fos mRNA. Consistent with this idea, DOR1 signaling leading to c-Fos mRNA induction was insensitive to wortmannin (Fig. 8A).
The c-Fos protein can participate in forming the AP-1 transcription factor complex (21,58); therefore, we examined the regulation of AP-1 in response to DOR1 signaling. Fig. 8B shows that DOR1 signaling elevated levels of DNA-binding AP-1 complexes. Like DOR1-mediated activation of ERK and c-Fos mRNA induction, DOR1 stimulation of DNA-binding AP-1 complexes was sensitive to PD098059 but not wortmannin. In addition, we previously showed that like DOR1 activation of ERK, the DOR1-mediated increase in DNA-binding AP-1 is abrogated by PTX (35). The AP-1 complexes mobilized by DOR1 include those that are transcriptionally active, since DOR1 ligation also elevated transcription from an AP-1 reporter plasmid in a manner sensitive to PTX or PD098059 but not to wortmannin (Fig. 8, C and D). These results indicate that the novel ␤␥-, Ras-, and PI 3-kinase-independent pathway of ERK activation described above is essential for DOR1 signaling that leads to increased levels of c-Fos mRNA and transcriptionally active AP-1 transcription factors.
A Role for a Tyrosine Kinase in ERK Activation by DOR1/␣i Signaling-Several reports indicate that tyrosine kinases can participate in G i protein-mediated stimulation of ERK (6,7,(13)(14)(15)(16); therefore, we examined the effects on DOR1 signaling of a tyrosine kinase inhibitor, herbimycin A. Pretreatment with herbimycin A partially inhibited the activation of ERK in response to DOR1 signaling (Fig. 9A). Herbimycin A pretreatment also partially inhibited the increase in DNA-binding AP-1 transcription factors in response to deltorphin (Fig. 9B). In contrast to its partial inhibition of DOR1 effects, herbimycin A almost completely blocked ERK activation and AP-1 mobiliza-tion in response to pervanadate, an inhibitor of tyrosine phosphatases that activates Src family tyrosine kinases in Jurkat T cells (39). Densitometric analysis shows that the DOR1-mediated increases in ERK2 phosphorylation and AP-1 DNA binding activity were inhibited 50 and 70%, respectively, by herbimycin A. These results suggest that the ␤␥-independent, DOR1-initiated pathway that stimulates ERK and increases AP-1 transactivation is regulated by a herbimycin A-sensitive tyrosine kinase. DISCUSSION Signaling by PTX-sensitive G proteins leads to activation of ERK1 and ERK2 (1-16; reviewed in Ref. 17), events that are important for mobilizing AP-1 and for stimulating cell cycle entry in response to growth factors (17, 20 -22, 58). Most studies on the molecular mechanisms that underlie this response have focused on ERK activation by G i or G o protein ␤␥ subunits (reviewed in Ref. 17), yet recent evidence for ␤␥-independent pathways (10, 13) and the oncogenic potential of mutant G i protein ␣i subunits (24 -26) indicate that alternative mechanisms exist. Here, we present results that establish the existence in Jurkat T cells of distinct ␤␥-mediated and ␤␥-independent pathways for G i protein-mediated stimulation of MEK-1 and ERK activity. In addition, we present evidence that the ␤␥-independent pathway is independent of Ras and PI 3-kinase activity and demonstrate that it is required for Elk-1-dependent transcription and the mobilization of AP-1 transcription factors in response to signaling by a G i protein-coupled receptor.
We analyzed the molecular mechanisms of G i protein-mediated ERK stimulation using the DOR1/Ju.1 subline of the human T lymphocyte cell line, Jurkat. DOR1/Ju.1 cells are stably transfected with the neuronal G i protein-coupled receptor, DOR1, which we previously showed signals via a PTXsensitive G i protein in these cells (35,36). Here, we show that agonist stimulation of DOR1 increases ERK activity and Elk-1-dependent transcriptional activity via a PTX-sensitive pathway, establishing that this pathway requires a G i protein. We further show that these effects of DOR1 are sensitive to the MEK inhibitor, PD098059, and that DOR1 activates MEK-1, indicating that this PTX-sensitive pathway activates ERK and downstream transcriptional events via MEK-1.
Since both MEK-1 activity and Elk-1-dependent transcription were elevated by transiently transfecting G i protein ␤ 1 and ␥ 2 subunits into Jurkat T cells, we asked if ␤␥ mediates DOR1 coupling to ERK. Interestingly, ␤ARKct inhibited Elk-1-dependent transcriptional activity in response to overexpressed ␤␥ but had no effect on Elk-1-dependent transcription that followed ligation of DOR1. Similarly, ␤ARKct inhibited constitutive MEK-1 activation by overexpressed ␤␥ but had no effect on MEK-1 activation in response to DOR1 signaling. In other cell types, ␤ARKct inhibits ERK activation in response to ligation of certain G i protein-coupled receptors, apparently because ␤ARKct sequesters free ␤␥ (2, 3, 10). Our results therefore indicate that although signaling by ␤␥ subunits can stimulate MEK-1 and downstream events in Jurkat T cells, the G i protein-coupled receptor, DOR1, stimulates MEK-1-, ERK-, and Elk-1-dependent transcription via a ␤␥-independent pathway.
We next asked if signaling intermediates that generally participate in ERK regulation also participate in the ␤␥-independent pathway. G protein ␤␥ subunits (1)(2)(3)22), as well as receptors that signal using tyrosine kinases (22,49), stimulate ERK via activation of Ras, a GTP/GDP-binding oncoprotein that stimulates the Raf/MEK/ERK kinase cascade. Although the DOR1-and ␤␥-mediated pathways were each sensitive to the MEK inhibitor, PD098059, both ERK activity and Elk-1-dependent transcription in response to DOR1 ligation were in- sensitive to transient expression of the dominant-negative Ras, RasN17. In contrast, RasN17 inhibited Elk-1-dependent transcription in response to either ␤␥ overexpression or TCR-CD3 ligation, consistent with previous reports that these stimuli use Ras to activate ERK (1-3, 11, 49, 50). We also explored the role of PI 3-kinase activity in the ␤␥-independent pathway. Inter-estingly, DOR1 stimulation of ERK or Elk-1-dependent transcription was insensitive to the PI 3-kinase inhibitor, wortmannin, and did not synergize with constitutively elevated PI 3-kinase activity that enhanced ␤␥ stimulation of Elk-1-dependent transcription.
Together, these results indicate that DOR1 activates deltorphin for 4 h, and nuclear extracts were prepared and assayed for AP-1 by electromobility shift assay using a consensus AP-1 binding site probe. Control, assay performed in the presence of 100-fold excess unlabeled probe, No Ex., assay performed in the absence of nuclear extract. C, cells were transfected with a plasmid containing an AP-1-responsive promoter upstream of a luciferase gene, plated in 0.5% serum-containing medium for 24 h, pretreated 30 min with 100 nM wortmannin where indicated, stimulated as indicated for 5 h, and assayed for luciferase activity as in Fig. 1. Results shown are from a representative experiment performed in duplicate; similar results were obtained in three independent experiments. D, results of multiple experiments assaying AP-1 promoter activity performed as in C except that pretreatments were with PTX-B, PTX, or PD098059 (PD). Stimuli were 10 Ϫ7 M deltorphin or 50 ng/ml PMA for 5 h. Each bar denotes the mean Ϯ S.E. of 3-11 independent determinations. *, significantly different from deltorphin or PMA alone (p Ͻ 0.001).
MEK-1-, ERK1-and ERK2-, and Elk-1-dependent transcriptional activity in Jurkat T cells via a mechanism that requires neither Ras nor a PI 3-kinase. In contrast, ␤␥-mediated stimulation of MEK-1-and Elk-1-dependent transcription in these cells is sensitive to RasN17 and synergizes with PI 3-kinase activation. Since DOR1, but not ␤␥, signaling is also insensitive to ␤ARKct, these results suggest that DOR1 and ␤␥ subunits stimulate ERK activity via independent signaling pathways that differ in their requirements for Ras. Since DOR1 stimulation of ERK was partially sensitive to the tyrosine kinase inhibitor, herbimycin A, the ␤␥-independent pathway may involve a tyrosine kinase. Experiments aimed at identifying the tyrosine kinase(s) associated with DOR1 signaling are in progress. However, it seems unlikely that tyrosine kinases associated with the TCR-CD3 receptor complex are involved via transactivation in DOR1 activation of ERK, since RasN17 inhibited Elk-1-dependent transcription in response to signaling by TCR-CD3 but not DOR1.
To our knowledge, this is the first evidence for a G i proteinmediated but ␤␥-independent pathway of ERK activation that is independent of both Ras and PI 3-kinase activity. Since DOR1 uses a PTX-sensitive mechanism to stimulate this pathway and since lymphoid cells do not express other ␣ subunits that are PTX targets (28), it is most likely that DOR1 uses uses ␣i2 and/or ␣i3, rather than ␤␥, to activate MEK-1, ERK, and downstream transcriptional events. Consistent with this idea, previous biochemical studies have shown that DOR1 can bind and promote the activation of G proteins containing both ␣i2 and ␣i3 subunits (Ref. 59 and references therein). The DOR1mediated pathway resembles the pathway that is stimulated by the ␣i2 oncogenic gip2 mutants, which activates ERK (25) and transforms Rat1a fibroblasts via RasN17-resistant pathways (24). However, unlike G i protein signaling that activates ERK via the epidermal growth factor receptor in COS-7 cells (13), DOR1 activates ERK in Jurkat cells via a pathway resistant to the PI 3-kinase inhibitor drug, wortmannin. A Ras-and ␤␥-independent pathway was recently described in Chinese hamster ovary epithelial cells, but this pathway appears to be mediated by ␣o subunits and is inhibited by down-regulation of PMA-sensitive protein kinase C isoforms (10). DOR1 signaling most likely stimulates ERK via a distinct pathway, since DOR1 signaling was not sensitive to protein kinase C inhibitor drugs that inhibit PMA activation of ERK, 2 and since lymphocytes express the G i protein subunits ␣i2 and ␣i3 but are deficient in other PTX-sensitive G proteins such as ␣o (28).
Interestingly, DOR1 signaling showed no evidence of initiating the ␤␥-mediated pathway of ERK activation, although receptor activation of the heterotrimeric G protein theoretically produces equal numbers of ␣i and ␤␥ signaling moieties. Other reports of putative ␣i/o-mediated ERK activation also do not address the question of why ␤␥-mediated signaling is not detected in these cases (10). It is possible, however, that the molecules required for ␤␥-mediated ERK activation, including the p110-␥ PI 3-kinase and Ras, are inaccessible to ␤␥ that is produced in the vicinity of the activated DOR1 receptor. It is also possible that free ␤␥ is itself efficiently sequestered when it is produced in membrane subdomains associated with specific receptors. Further research will be necessary to adequately address the question of what happens to the free ␤␥ subunits during ␤␥-independent signaling by DOR1 in Jurkat cells.
Although constitutively active ␣i2 subunits (␣i2Q205L) have been shown to activate ERK in Rat1a fibroblasts (24,25), we observed only a small elevation of Elk-1-dependent transcription and MEK-1 activity following transient transfection of such mutant ␣i2 or ␣i3 subunits into Jurkat T cells. 2 This may indicate that the ␣i-mediated ERK activation pathway in these cells is rapidly down-regulated in response to chronic stimulation. An alternative explanation for our results is that ␣i stimulates the ␤␥-independent pathway by acting together with a second DOR1-activated ␣ subunit. Biochemical studies show that besides PTX-sensitive ␣i and ␣o subunits, DOR1 is capable of physically interacting with PTX-insensitive G proteins that contain ␣ q , ␣ z/x , and ␣ 16 (59). However, it is unlikely that DOR1 signaling in Jurkat T cells activates these ␣ subunits in addition to ␣i. Signaling by ␣q-type subunits (including ␣ 16 ) can directly activate PLC-␤ (60). In contrast, DOR1 signaling increases calcium in Jurkat T cells via a mechanism that is fully sensitive to PTX (36). Similarly, the predominantly brain-expressed ␣ z/x subunit permits adenylyl cyclase regulation in a PTX-insensitive manner, while DOR1 uses a PTX-sensitive mechanism to decrease cAMP levels in Jurkat T cells (36). While we are directly addressing this question by identifying the G proteins that co-purify with DOR1 in Jurkat T cells, we consider it most likely that the ␤␥-independent ERK activation pathway we describe here is mediated by ␣i2 and/or ␣i3.
Finally, we show that the ␤␥-independent pathway for ERK activation is required for DOR1 to mobilize AP-1 transcription factors in the Jurkat T cell line, a result indicating that this novel ␤␥-independent pathway contributes to the regulation of a downstream event relevant to mitogenic signaling. DOR1 signaling increased mRNA for the AP-1 subunit, c-Fos, and levels of transcriptionally active AP-1, via mechanisms sensitive to a MEK inhibitor drug that also blocked DOR1 stimulation of ERK. Moreover, neither c-Fos transcription nor AP-1 mobilization in response to DOR1 signaling was sensitive to the PI 3-kinase inhibitor, wortmannin. We previously showed that DOR1 increases AP-1 in these cells via a PTX-sensitive 2 K. E. Hedin, unpublished observations. FIG. 9. DOR1 signaling stimulates ERK and mobilizes AP-1 via a pathway sensitive to herbimycin A. DOR1/Ju.1 Jurkat T cells were pretreated for 16 h with 3 M herbimycin A as indicated. A, cells were then stimulated for 5 min with 10 Ϫ7 M deltorphin (Delt) or 0.1 mM pervanadate (Perv.) and assayed for active, Thr 202 /Tyr 204 -phosphorylated ERK1 and ERK2 as in Fig. 1A. B, cells were stimulated with 10 Ϫ7 M deltorphin or 0.1 mM pervanadate for 4 h, and nuclear extracts were prepared and assayed for AP-1 by EMSA as in Fig. 8B. pathway (35). Together with signals from TCR-CD3 and CD28, the DOR1-mobilized AP-1 complexes enhance the transcriptional activity of the NF-AT/AP-1-binding element of the interleukin-2 (IL-2) promoter and increase IL-2 synthesis and secretion (35). Others have also shown that G i protein-coupled receptors enhance the secretion of IL-2 (33,34). IL-2 secretion occurs coincident with T lymphocyte immune activation and cell cycle entry and acts on T cells to promote their growth and continued immune activation. The ␤␥-independent pathway of ERK activation may, therefore, form part of a mechanism that permits G i protein-coupled receptors to regulate IL-2 and other AP-1-responsive genes in T lymphoid cell types.