Introduction

Top Summary Introduction Materials & Methods Results Discussion References

The GH₃ rat anterior pituitary cell line often is used as a model for lactotrophs because it produces prolactin and possesses functional receptors for TRH and other regulatory neuropeptides (e.g., vasoactive intestinal peptide or bombesin) that act to increase prolactin secretion [see Ben-Jonathan et al. (1989) and Corette et al. (1995) for reviews]. Both early releases of preformed prolactin and sustained secretion of newly synthesized hormone contribute to TRH effects (Gershengorn, 1986). Activation of the GH₃ cell TRH-R promotes its interaction with one or more G proteins with subsequent regulation of several effectors, including PLC and different ionic channels (Corette et al., 1995; Gershengorn and Osman, 1996). As for many cell types and agonists, phosphorylation of proteins found in GH₃ cell cytosolic, microsomal, and nuclear fractions has been repeatedly reported for TRH (Drust et al., 1982; Drust and Martin, 1982; Sobel and Tashjian, 1983; Drust and Martin, 1984; Jefferson et al., 1991), and phosphorylation-dephosphorylation cycle or cycles have been implicated in regulation of GH₃ cell electrical activity by modulation of specific ionic currents (Barros et al., 1992, 1993; Delgado et al., 1992). Among other protein kinases (e.g., PKC and Ca²⁺/calmodulin kinase; Drust and Martin, 1984; Martin et al., 1990; Jefferson et al., 1991), MAPK has been shown recently to be stimulated by TRH in GH₃ cells (Ohmichi et al., 1994). Although partially dependent on PKC activity, a PKC-independent pathway also was found to couple the TRH-R to MAPK activation; however, the molecular entities involved in such a coupling remain obscure. The physiological role of this TRH effect is not known, but early tyrosine phosphorylation leading to MAPK activation has been suggested to be involved in TRH-induced prolactin secretion (Kanda et al., 1994; Ohmichi et al., 1994). Furthermore, the known regulatory role of the enzyme on gene transcription makes it a candidate to participate in prolactin synthesis or proliferative and neuronal trophic responses also reported for TRH (Metcalf and Jackson, 1989).

Cloning of the GH₃ cell TRH-R allowed the deduction of its amino acid sequence and its inclusion in the superfamily of seven transmembrane-spanning G protein-coupled receptors (Gershengorn and Osman, 1996). Activation of the ubiquitous MAPK by growth factor receptors that possess intrinsic tyrosine kinase activity has been largely recognized. Mitogenic signals from these receptors involve a series of SH₂- and SH₃-dependent protein/protein interactions resulting in p21^ras-dependent MAPK activation [see Malarkey et al. (1995) and Marshall (1995) for reviews]. Recently, a number of receptors that couple to heterotrimeric G proteins have been shown to stimulate MAPK activity (Crespo et al., 1994a; Faure et al., 1994; Russell et al., 1994; Hawes et al., 1995; Malarkey et al., 1995; Van Biesen et al., 1995; Luttrell et al., 1996), including both receptors that couple to G_q and G_i. It has been suggested that MAPK activation in COS cells via receptors coupled to members of the G_q/11 family largely occurs through a PKC-dependent pathway. In this case, the GTP-bound  subunit of G_q/11 would activate PLC, which also would activate MAPK via a mechanism not completely understood (Faure et al., 1994; Hawes et al., 1995; Luttrell et al., 1996). In contrast, MAPK activation by receptors coupled to PTX-sensitive G proteins (e.g., G_i or G_o) occurs via a pathway PKC independent, p21^ras mediated, and dissociable from PLC activation. dimers released from the PTX-sensitive G protein and one or more tyrosine phosphorylations are important components of the coupling cascade to MAPK initiated in this type of receptors (Crespo et al., 1994a; Faure et al., 1994; Koch et al., 1994; Hawes et al., 1995; Van Biesen et al., 1995; Luttrell et al., 1996). However, it has been reported that stimulation of MAPK activity by G_q-coupled m1-R is partially independent of PKC activity (Crespo et al., 1994b), and some inhibitory effect of overexpression of transducin  subunit on m1-R-induced MAPK activation also has been obtained (Crespo et al., 1994a). This opens the possibility that a PKC-independent dimer-mediated route also is involved, at least in part, in MAPK activation via receptors coupled to PTX-insensitive G proteins.

In this report, the transduction pathways linking TRH-R stimulation to MAPK activation are explored using COS-7 cells transfected with the TRH-R and an epitope-tagged MAPK. Our results indicate that as in pituitary cells, TRH-Rs activate MAPK via a pathway only partially dependent on PKC activity. Furthermore, they show that dimers of a PTX- and CTX-insensitive G protein are involved in the PKC-independent fraction of the dual signaling route to MAPK initiated in the TRH receptor.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Chemicals. TRH, carbachol, PMA, MBP, genistein, PTX, CTX, and culture medium were purchased from Sigma Chemie (Deisenhofen, Germany). Serum was from Biochrom (Berlin, Germany). Anti-hemagglutinin monoclonal antibody 12CA5 and human EGF was from Boehringer-Mannheim Biochemica (Mannheim, Germany). [-³²P] ATP (7000 Ci/mmol, 160 mCi/ml) and myo-[³H]inositol (17.6 Ci/mmol, 1 Ci/ml) were obtained from Amersham International (Buckinghamshire, UK). [³H]Met-TRH (pGlu-3-methyl-His-ProNH₂; 82.5 Ci/mmol, 1 mCi/ml) was from New England Nuclear (Boston, MA). GF109203X was from Calbiochem (San Diego, CA). DEAE-dextran and Gamma-Bind Sepharose beads were from Pharmacia Biotech (Uppsala, Sweden) (St. Quentin, France). All other reagents were purchased from Sigma and were the highest quality available.

Expression plasmids. The expression plasmid pcDNA3-HA-MAPK (provided by Dr. J. S. Gutkind, National Institute of Dental Research, National Institutes of Health, Bethesda, MD) contains an amino-terminal hemagglutinin-tagged murine p42-MAPK cDNA. Its protein product (referred to here as HA-MAPK) can be recognized efficiently by the murine monoclonal antibody 12CA5. The previously described cDNAs for rat TRH receptor isoforms (de la Peña et al., 1992) were cloned in pcDNA3 (InVitrogen, San Diego, CA) as HindIII/EcoRI fragments. m1-R and m2-R cDNAs were subcloned in pcDNA3 as BamHI/XbaI and XhoI/XbaI fragments, respectively, transferred from constructs in pBluescript KS⁺ (Stratagene, La Jolla, CA). G_t was cloned in pcDNA3 as an EcoRI/XhoI fragment transferred from pcDNA I. The plasmid constructs of wild-type or mutationally active G protein  subunits cloned in pcDNA I and that of p21^N17ras cloned in pcDNA 3 also were generously provided by Dr. J. S. Gutkind.

Transient expression in COS-7 cells. COS-7 cells were cultured in Dulbecco's modified Eagle's medium/Ham's F-12 (1:1 mixture) supplemented with 0.13% NaHCO₃, 50 IU/ml penicillin, 50 µg/ml streptomycin, and 10% fetal bovine serum. Plasmid DNA transfection of subconfluent COS-7 was performed by the DEAE-dextran technique as described previously (del Camino et al., 1997).

MAPK assay. MAPK activity in COS-7 cells transfected with HA-MAPK was determined as described previously (Crespo et al., 1994a). COS-7 cells were cotransfected with the different DNA constructs (1-2 µg/plate) and the expression plasmid pcDNA 3-HA-MAPK (1-2 µg/plate). The total amount of plasmid DNA was equalized in all 60-mm plates with vector DNA (pcDNA3) when necessary. Two days after transfection, cells were serum-fasted overnight and then stimulated with the different agents. After the indicated periods of time, cells were washed with cold phosphate-buffered saline and lysed in a buffer containing 10 mM EGTA, 40 mM -glycerophosphate, 1% Triton X-100, 2.5 mM MgCl₂, 2 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml aprotinin, 10 µg/ml leupeptin, and 20 mM HEPES, pH 7.5. After centrifugation, clarified supernatants were immunoprecipitated with an anti-hemagglutinin monoclonal antibody 12CA5 for 1 hr at 4°. Immunocomplexes were recovered with the use of Gamma-Bind Sepharose beads. Pellets of individual plates were washed three times with phosphate-buffered saline solution supplemented with 1% Triton X-100; once with 0.5 M LiCl in 100 mM Tris, pH 7.5; and once with kinase reaction buffer [12.5 mM -glycerophosphate, 7.5 mM MgCl₂, 0.5 mM EGTA, 0.5 mM sodium fluoride, 0.5 mM sodium vanadate, 12.5 mM MOPS, pH 7.5]. Reactions were performed at 30° for 30 min by the addition to pellets of a 30-µl volume of kinase reaction buffer containing 1 µCi of [-³²P]ATP, 20 µM unlabeled ATP, and 1.5 mg/ml MBP. Reactions were terminated by the addition of 5× Laemmli's buffer, boiled, and electrophoresed on 15% polyacrylamide gels. Phosphorylated MBP was visualized by autoradiography and quantified with the aid of an InstantImager (Packard, Meriden, CT).

Measurement of [³H]Met-TRH binding and phosphatidylinositol hydrolysis in transfected cells. Binding assays to COS-7 cells were performed as described previously (del Camino et al., 1997). For assays of inositol phosphate accumulation, cells were incubated 48 hr after transfection with 1 µCi/ml myo-[³H]inositol for 24 hr at 37° in inositol-free medium supplemented with 5% fetal bovine serum. Cells were incubated for an additional 6 hr in serum-free conditions and stimulated with experimental agents for 45 min in the presence of 10 mM LiCl. Inositol phosphates were extracted subsequently and analyzed by ion exchange chromatography using AG1-X8 resin (BioRad, Hercules, CA).

Statistical analysis. Unless otherwise indicated, data are expressed in the histograms as mean ± standard error for the number of experiments indicated, which were performed with duplicate or triplicate samples. Due to significant day-to-day differences in MAPK activation levels, only data obtained on the same day with different treatments were used for comparison. Fold stimulation in the graphs corresponds to MAPK activity levels compared with those obtained in untreated cells. Basal activity without any treatment has not been subtracted in any case. Thus, 1-fold activation in the presence of an agonist plus an inhibitor would correspond to a 100% inhibition of the agonist-induced activation.

	Results

Top Summary Introduction Materials & Methods Results Discussion References

The addition of TRH to COS-7 cells cotransfected with rat GH₃ cells TRH-R and an epitope-tagged MAPK caused a time- and concentration-dependent activation of the enzyme (Fig. 1). The increase in activity measured 5 min after the start of 1 µM TRH treatment averaged 7.2 ± 0.6-fold (47 experiments) of control values without hormone. Although less marked than the activation reached by stimulation of COS-7 EGF endogenous receptors with EGF (22.3 ± 3.8-fold, 21 experiments; see Fig. 1A), these results demonstrate that the TRH-R is able to activate MAPK in transfected cells, as demonstrated previously in GH₃ anterior pituitary cells (Ohmichi et al., 1994). Both isoforms of the receptor (TRH-R₄₁₂ and TRH-R₃₈₇; de la Peña et al., 1992) induced equal activations of the kinase (Fig. 1A) and showed a similar dependence on hormone concentration (not shown). Thus, subsequent experiments were performed routinely with the TRH-R₄₁₂ long isoform. The MAPK activation in response to TRH was rapid, reaching a maximum after 5 min of exposure to the neuropeptide that slightly declined thereafter up to 30 min of treatment (Fig. 1, B and D). The activation level at 5 min after stimulation depended on the concentration of TRH (Fig. 1, C and D). The EC₅₀ value was obtained at 1-10 nM, a value similar to the K_d value for hormone-receptor interactions in GH₃ cells (Hinkle, 1989), and maximal stimulation levels were obtained at ~100 nM TRH.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Stimulation of MAPK in COS-7 cells after stimulation of transfected TRH-R. A, TRH increases the enzymatic activity of an epitope-tagged MAPK when cotransfected with both isoforms of the TRH-R. COS-7 cells were cotransfected with pcDNA 3 plasmid DNA (vector) or with the same plasmid containing the sequences encoding the long (TRH-R412) or the short (TRH-R387) isoforms (de la Peña et al., 1992) of the TRH-R and HA-MAPK as described in Materials and Methods. MAPK activity was determined in serum-fasted cells after a 30-min treatment with (TRH) or without () 1 µM TRH. MAPK activity in response to stimulation of COS-7 cell endogenous EGF receptors with 100 ng/ml EGF is shown for comparison. Radioactivity incorporated in MBP as a result of the enzymatic activity of immunoprecipitated HA-MAPK was visualized by autoradiography. B, Time course of MAPK activation by TRH in TRH-R-transfected COS-7 cells. Incorporation of radioactivity in MBP was measured in reactions performed as in A after stimulation of cells cotransfected with TRH-R412 and HA-MAPK. Cells were stimulated with 1 µM TRH for the indicated periods of time. A representation of the amount of radioactivity incorporated in MBP as a function of the duration of TRH treatment is shown in D (left). C, Dependence of MAPK activation on TRH concentration. MAPK activity was measured in HA-MAPK immunoprecipitates of COS-7 cells expressing TRH-R and stimulated with the indicated TRH concentrations for 5 min. The representation of the amount of radioactivity incorporated in MBP as a function of the TRH concentration is shown in D (right). A-C, Major bands correspond to the position of MBP in the autoradiograms.

It has been shown previously that both G_i- and G_q/11-coupled receptors are able to activate MAPK in COS cells (Crespo et al., 1994a; Faure et al., 1994; Russell et al., 1994; Hawes et al., 1995; Van Biesen et al., 1995; Luttrell et al., 1996). It also is known that TRH-Rs are coupled to PTX- and CTX-insensitive G_q/11 for activation of PLC in GH₃ cells and to other G proteins for other effects (Gershengorn and Osman, 1996). As an initial approach to determine the coupling specificity of TRH-R to MAPK activation, we used bacterial toxins for covalent modification of COS-7 G proteins. As shown in Fig. 2, the response to TRH was not altered by pretreatment of the cells with PTX or CTX. Failure to detect toxin effects was not due to lack of modifications of cellular G proteins by the toxins, because MAPK activation mediated by m2-R was 83% inhibited by PTX pretreatment, as reported previously (Crespo et al., 1994a).

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Effect of PTX and CTX treatment on MAPK activation caused by stimulation of TRH-R and m2-R in receptor-transfected COS-7 cells. COS-7 cells were transiently transfected with DNA plasmids encoding HA-MAPK and TRH-R or m2-R as described in Materials and Methods. When indicated, cells were preincubated overnight with PTX (100 ng/ml) or CTX (200 ng/ml) before stimulation with 1 µM TRH (TRH-R) or 1 mM carbachol (m2-R) for 5 min. A, Autoradiogram showing the radioactivity incorporated in MBP. Only the representative result of a single reaction from triplicate samples is shown for every individual treatment. B, Averaged results of four experiments performed in triplicate. Bars, averaged MAPK activity expressed as fold stimulation respective to that obtained without TRH or carbachol for TRH-R- and m2-R-transfected cells, respectively. Significantly lower MAPK activation values were obtained after the addition of carbachol to cells expressing m2-Rs and incubated with PTX (p < 0.05 versus untreated cells, paired t test).

TRH-R stimulation in anterior pituitary cells causes G_q/11-mediated PIP₂ hydrolysis and PKC activation (Gershengorn and Osman, 1996). However, both PKC-dependent and -independent pathways are involved in GH₃ cell MAPK activation (Ohmichi et al., 1994). Previous work in COS-7 cells has demonstrated two main pathways for G protein-mediated MAPK activation. G_i-coupled receptor-mediated MAPK activation is G subunit mediated, p21^ras dependent, PKC independent, and dissociable from PLC activation (Crespo et al., 1994a; Faure et al., 1994; Koch et al., 1994; Hawes et al., 1995; Van Biesen et al., 1995; Luttrell et al., 1996). The mechanism or mechanisms leading to MAPK activation by G_q-coupled receptors, such as m1-R and _1B-adrenergic receptor, remain controversial. Thus, a p21^ras-independent, PLC- and PKC-mediated MAPK activation pathway insensitive to G subunit sequestrants (e.g., overexpression of transducin  subunits or a peptide derived from the carboxyl terminus of the -adrenergic receptor kinase) has been involved in such a response (Faure et al., 1994; Koch et al., 1994; Hawes et al., 1995). However, stimulation of MAPK activity by m1-R only partially dependent on PKC and partial inhibitions of m1-R induced activations by transducin  subunits also have been reported (Crespo et al., 1994a, 1994b). To determine whether a PKC-independent pathway might be involved in TRH-induced activation of MAPK, we initially depleted COS-7 cells of PKC by a 18-hr treatment with 1 µM PMA. This prevented subsequent activation of MAPK by short term treatment with PMA and TRH. However, a several-fold MAPK activation was detected at the end of the prolonged PMA treatment, even without MAPK stimulants (data not shown). This might confound experimental interpretations. As an alternative approach, we used the nontoxic PKC-specific inhibitor bisindolylmaleimide GF109203X, previously shown to act as a potent in vivo PKC inhibitor in COS-7 cells (Crespo et al., 1994b; Coso et al., 1995). As shown in Fig. 3, treatment of the cells with 1 µM GF109203X almost completely abolished the PMA-stimulated MAPK activity. Furthermore, no detectable effect of the inhibitor was obtained on the PKC-independent activation of the enzyme by EGF. Finally, only a partial inhibition of ~50% of the TRH-induced activation of MAPK was obtained in the presence of GF109203X. This suggests that as in pituitary cells, MAPK activation in transfected COS-7 cells is only partially dependent on PKC.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Effect of the PKC inhibitor GF109203X on TRH-, PMA-, and EGF-mediated MAPK activation. COS-7 cells were cotransfected with plasmid DNA encoding HA-MAPK and stimulated for 5 min with 1 µM TRH or 50 ng/ml PMA or for 3 min with 100 ng/ml EGF as indicated. Plasmid-encoding TRH-R was included for transfection of TRH-stimulated cells. Before stimulation, cells were preincubated with 3.5 µM GF109203X for 60 min. The inhibitor was present during treatment with different agonists. A, Autoradiogram of the radioactivity incorporated in MBP. B, Averaged results for the number of experiments indicated (N) and performed in triplicate for TRH-treated (left) or PMA-treated (right) cells. GF, presence of GF109203X. MAPK activity is expressed as fold stimulation respective to that obtained without agonists. Treatment with GF109203X significantly reduced MAPK activation by TRH or PMA (p < 0.01 versus untreated cells, paired t test).

As stated, PKC-dependent activation of MAPK in COS-7 cells involves a pathway independent of p21^ras (Crespo et al., 1994a, 1994b; Russell et al., 1994; Hawes et al., 1995; Luttrell et al., 1996). Further support for the existence of a PKC-independent fraction of TRH-induced MAPK activation was obtained in experiments in which p21^ras function is impaired by expression of the dominant negative mutant p21^N17ras. As shown in Fig. 4A, little effect of p21^N17ras expression is detected on PKC-dependent MAPK activation by PMA. However, a 51% inhibition of TRH-induced MAPK activity is exerted by introduction of p21^N17ras into the cells. It is interesting that as for the GF109203X effect and consistent with the coexistence of PKC-dependent and -independent pathways going to MAPK from TRH-R, only a partial inhibition of the TRH-induced enzyme activation was obtained in the presence of p21^N17ras. This partial reduction of TRH-induced activation was not due to a lowered potency of p21^N17ras to inhibit Ras signaling. Thus, a similar 60% inhibition of the TRH-evoked response was obtained in a second set of experiments in which the Ras-dependent m2-R response (Crespo et al., 1994a; Russell et al., 1994) was almost abolished (87% inhibition) by coexpression of p21^N17ras (Fig. 4B). Finally, the inhibitory effect of p21^N17ras was not related to an indirect effect on TRH-R expression, because similar amounts of [³H]Met-TRH binding sites were obtained in cells transfected with TRH-R alone (20 × 10³ receptors/cell, mean of three determinations) or with both TRH-R and p21^N17ras (25 × 10³ receptors/cell), as measured by [³H]Met-TRH saturation binding. These results demonstrate that p21^N17ras constitutes a useful tool to specifically distinguish between pure PKC-dependent responses, such as that promoted by PMA, and PKC-independent Ras-mediated activations, such as that induced by m2-R. Furthermore, they indicate that a PKC-dependent route coexists with a Ras-dependent pathway leading to MAPK activation in response to TRH.

View larger version (33K): [in this window] [in a new window]   Fig. 4.   Effect of expression of the dominant negative form of p21ras (p21N17ras) on MAPK activation. A, Expression of p21N17ras partially inhibits TRH-mediated, but not PMA-mediated, MAPK activation. COS-7 cells were cotransfected with DNA plasmids encoding HA-MAPK and TRH-R plus either vector DNA alone (filled bars) or plasmid DNA encoding p21N17ras (striped bars). Plasmid DNA containing TRH-R sequence was not used for PMA-treated cells. Cells were stimulated with 1 µM TRH or 50 ng/ml PMA for 5 min. B, Expression of p21N17ras inhibits TRH- and carbachol-mediated MAPK activation differently in cells expressing TRH-R and m2-R, respectively. COS-7 cells were cotransfected with DNA plasmids encoding HA-MAPK and TRH-R or m2-R plus either vector DNA alone (filled bars) or plasmid DNA encoding p21N17ras (striped bars). Cells were stimulated with 1 µM TRH (TRH-R) or 1 mM carbachol (m2-R) for 5 min. MAPK activity is expressed as fold stimulation respective to that obtained without agonists. Values shown represent mean ± standard error from three separate experiments performed in triplicate. Different batches of cells were used to obtain the data. Except for PMA-induced activations, significantly lower MAPK activation values were obtained in cells cotransfected with p21N17ras (p < 0.05 versus non-p21N17ras-containing cells, paired t test).

The demonstration of both a PKC-dependent and -independent way to activate MAPK suggests the existence of a dual pathway initiated in the TRH-R. However, only partial inhibitions are induced by agents presumably acting on only one of both routes. This does not exclude that the action of these agents is exerted at different levels of a single route. To demonstrate that the TRH effect is indeed due to operation of two independent pathways converging at a point downstream of the step inhibited by GF109203X and p21^N17ras, additivity of inhibitions caused by these agents was studied. As shown in Fig. 5, both GF109203X and p21^N17ras caused a partial inhibition of MAPK activation by TRH, which amounted to 67% and 75%, respectively, in this set of experiments. Furthermore, the TRH-induced activity was completely abolished (inhibition level of 97%) when both agents were used at the same time. This indicates that two routes couple TRH-R to MAPK activation: a PKC-dependent pathway, probably involving a G_q/11 protein coupled to phospholipase C activation and generation of diacylglycerol, and a second route that involves p21^ras and activates MAPK independent of PKC activity increases.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Additivity of GF109203X and p21N17ras inhibitory effects on TRH-induced MAPK activation. COS-7 cells were cotransfected with DNA plasmids encoding HA-MAPK and TRH-R. When indicated, plasmid encoding p21N17ras was included for transfection (RasN17). Performance of a 60-min preincubation with 3.5 µM GF109203X before 5-min stimulation of the cells with 1 µM TRH is indicated. In this case, the PKC inhibitor was maintained during the TRH treatment. A, Autoradiogram showing the radioactivity incorporated into MBP. B, Averaged results of four experiments performed with triplicate samples. MAPK activity is expressed as fold stimulation respective to that obtained in the absence of TRH. MAPK activation values were significantly lower in the presence of GF109203X, p21N17ras, or both (p < 0.05 versus control, paired t test).

Previous work in GH₃ cells indicated the possible participation of tyrosine phosphorylation or phosphorylations in the PKC-independent activation of MAPK by TRH (Ohmichi et al., 1994). However, the mechanism or mechanisms leading to tyrosine kinase activity increases, or the nature of the tyrosine kinase or kinases implicated in the effect of the neuropeptide is not known. In COS-7 cells, the p21^ras-mediated and PKC-independent effects of G_i protein-coupled receptors on MAPK are exerted by liberation of dimers and subsequent stimulation of p21^ras, with participation of one or more tyrosine kinases (Hawes et al., 1995; Van Biesen et al., 1995; Luttrell et al., 1996). To determine the role of free G protein dimers in MAPK activation by TRH, we used the overexpression of t to sequester subunits after they are released from G proteins by receptor stimulation (Crespo et al., 1994a; Faure et al., 1994). The presence of t did not modify significantly the amount of expressed TRH-Rs, as measured by [³H]Met-TRH saturation binding (40 × 10³ receptors/cell in the absence versus 33 × 10³ receptors/cell in the presence of t, mean of three determinations). As shown in Fig. 6, expression of t did not affect the activation of MAPK by PMA but inhibited 77% of its activation when mediated by m2 muscarinic receptors. This result parallels those previously obtained by others, demonstrating that the activation of MAPK by m2-R is mediated predominantly via dimers released from G_i (Crespo et al., 1994a; Faure et al., 1994). Furthermore, TRH-induced activation of the enzyme via TRH-R was inhibited in the presence of t. However, consistent with coexistence of a PKC-dependent and a subunit-mediated activation initiated in the TRH-R, the inhibition of the TRH response by t was reduced to a 52%. It is important to note that this effect of t overexpression is specifically exerted on subunit-mediated signals and is not due to a nonspecific impairment of G protein function by chronically sequestering dimers. Thus, the inositol phosphates produced in response to TRH were 1225 ± 360 cpm/mg of protein in COS-7 cells transfected with the TRH-R and 1056 ± 125 cpm/mg of protein when the same batch of cells was cotransfected with the receptor and t. This demonstrates that the q/11-mediated stimulation of PIP₂ hydrolysis in response to TRH remained unaffected in the presence of t.

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Effect of expression of t on MAPK activation induced by stimulation of TRH-R and m2-R. COS-7 cells were cotransfected with DNA plasmids encoding HA-MAPK and TRH-R or m2-R as indicated, plus either vector alone (filled bars) or plasmid DNA encoding t (striped bars). Plasmid-containing receptor sequences was not used for PMA-treated cells. Cells were stimulated for 5 min with 1 µM TRH (TRH; TRH-R-transfected cells), 1 mM carbachol (CCh; m2-R-transfected cells), or 50 ng/ml PMA. A, Autoradiogram showing the radioactivity incorporated into MBP. B, Averaged results for the number of experiments indicated (N) and performed in triplicate for TRH- or carbachol-stimulated cells. MAPK activity is expressed as fold stimulation respective to that obtained in the absence of agonists. Significantly lower MAPK activation values for TRH and carbachol were obtained in the presence of t (p < 0.01 versus controls, paired t test).

The existence of two independent routes to MAPK from the TRH-R also would predict the combined effect of a PKC inhibitor (GF109203X) and a subunit scavenger (_t) to be additive. On the other hand, it could be expected that blockade of the Ras-dependent pathway by p21^N17ras is not enhanced in the presence of the sequestrant t, which presumably acts at a step of the same route closer to the TRH-R. Fig. 7 shows that both predictions are indeed correct. As shown in Fig. 7A, the inhibition of MAPK activation by TRH was 41% and 44% in cells cotransfected with t or incubated with GF109203X, respectively. Such an inhibition was increased to 81% when the two treatments were performed simultaneously with the same batch of cells. In contrast, although 35% and 65% inhibition was obtained in the presence of t and p21^N17ras, respectively, no further increase respect to the inhibition caused by p21^N17ras was observed in the presence of both p21^N17ras and t (Fig. 7B). These data confirm the hypothesis of a dual pathway, which is consistent with a PLC- and PKC-dependent route plus a PKC-independent route mediated by dimers leading to Ras activation.

View larger version (37K): [in this window] [in a new window]   Fig. 7.   Inhibitory effect of t is additive with that of GF109203X (A) but not with inhibition induced by expression of p21N17ras (B). COS-7 cells were cotransfected with DNA plasmids encoding HA-MAPK and TRH-R, plus either vector alone (filled bars and GF) or plasmid DNA-encoding Gt with or without p21N17ras (RasN17). When indicated, cells also were incubated with 3.5 µM GF109203X for 60 min before the addition of TRH (GF). Cells were stimulated for 5 min with 1 µM TRH. Averaged results are shown for the indicated number of experiments performed with triplicate samples. Different batches of cells were used to obtain the data. MAPK activity is expressed as fold stimulation respect to that obtained in the absence of TRH. Significantly lower MAPK activation values were obtained in the presence of GF109203X, p21N17ras, or t, either alone or combined (p < 0.05 versus control, paired t test).

The dimer-mediated activation of MAPK by different receptors in COS cells has been always associated with liberation of these subunits from PTX-sensitive receptor-coupled G proteins. In addition, expression of specific sets of dimers has been shown repeatedly to induce MAPK activation in these cells (Crespo et al., 1994a; Faure et al., 1994; Hawes et al., 1995; Van Biesen et al., 1995). However, our results demonstrate that the TRH-induced activation is completely insensitive to treatment of the cells with bacterial toxins. To gain some insights into the nature of the G protein coupling of TRH-Rs to MAPK activation, several wild-type and -QLs were transfected with or without the TRH-R. Initially, we expected, as for exogenously introduced subunits, that the presence of mutationally activated forms of the G protein or proteins coupled to the TRH receptor would mimic the TRH-induced activation.

In the absence of receptors, wild-type q or 13 subunits did not induce a significant activation of MAPK (not shown). Furthermore, although a modest enhancement of MAPK activity was observed sporadically in the presence of _q-QL, no consistent activation of the kinase was induced by transfecting the GTPase-deficient mutationally activated  subunits of G_i2, G_s, and G₁₃ (data not shown) as well as of G_q (Fig. 8). Previous experiments with activated mutants of G protein  subunits in COS cells yielded conflicting results. Thus, although a modest enhancement of MAPK activity has been reported in response to activated q and s (Faure et al., 1994), a lack of activation on expression of q-QL (Qian et al., 1993; Crespo et al., 1994a) or s-QL (Crespo et al., 1994a, 1995) also has been shown. The lack of activation in the presence of q-QL does not seem to be coherent with the demonstrated ability of this constitutively active subunit to persistently activate COS cell PLC, causing a sustained high steady state level of inositol phosphates (Qian et al., 1993, Crespo et al., 1994). Thus, elevated diacylglycerol levels could be expected under these conditions with a concomitant PKC-dependent elevation of MAPK activity. A possible explanation for these paradoxical results can be obtained from experiments in which MAPK activation by different agents is studied in the presence of q-QL. As shown in Fig. 8, the TRH-induced activation of MAPK is strongly inhibited by q-QL expression. Similar results were obtained with the activated form of 13 but not with the wild-type form of G_q or with activated i2-QL or s-QL (not shown). This indicates that failure to activate MAPK by q-QL alone is not due to lack of expression of the subunit. Furthermore, it also demonstrates that one or more steps of the cascade leading to MAPK are inhibited in the presence of the activated q. Fig. 8 also shows that the inhibitory effect of q-QL coexpression is not exclusive for TRH-induced MAPK activation. Thus, a similar inhibitory effect is exerted on MAPK activation induced by the addition of carbachol to cells expressing another G_q-coupled receptor, such as the m1-R. Nevertheless, the receptor-independent and PKC-mediated activation caused by PMA also was inhibited by the constitutively active form of q. This result would be compatible with an inhibitory effect due to down-regulation of PKC after its chronic stimulation in response to sustained elevations of diacylglycerol. However, the presence of q-QL similarly inhibited the PKC-independent activations induced by stimulation of endogenous EGF (Fig. 8) or transfected m2 muscarinic receptors (not shown). Thus, although some involvement of PKC down-regulation cannot be completely excluded for pathways that use the kinase, our results indicate that the inhibitory effect of q-QL is exerted at a step downstream of that at which PKC-dependent and -independent routes converge.

View larger version (34K): [in this window] [in a new window]   Fig. 8.   Effect of expression of q-QL on MAPK activation induced by stimulation of TRH-R, m1-R, EGF-R, or PKC. COS-7 cells were cotransfected with DNA plasmids encoding HA-MAPK and TRH-R or m1-R as indicated, plus either vector alone or plasmids encoding q-QL. Plasmid-containing receptor sequences were not used for PMA- or EGF-treated cells. When indicated, cells were stimulated with 100 ng/ml EGF (EGF) for 3 min or with 1 µM TRH (TRH), 1 mM carbachol (CCh), or 50 ng/ml PMA (PMA) for 5 min. Averaged results are shown for the number of experiments indicated (N) and performed in triplicate. Note the absence of MAPK stimulation in agonist-untreated cells expressing the activated  subunit of Gq. In all cases, MAPK activation values were significantly lowered in the presence of q-QL (p < 0.05 versus controls, paired t test).

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

Recent work in GH₃ anterior pituitary cells demonstrated the ability of TRH-R stimulation to activate MAPK (Ohmichi et al., 1994). Coupling of TRH-R to bacterial toxin-insensitive q/11 protein activates a signaling cascade starting in PLC-mediated hydrolysis of PIP₂ that seems to be important for TRH-induced physiological effects (Gershengorn and Osman, 1996). However, PTX-sensitive G proteins are involved in TRH regulation of ionic channels in pituitary cells (Gollasch et al., 1993). Furthermore, CTX-sensitive G_s-like proteins have been implicated in modifications of cell excitability by the neuropeptide (Barros et al., 1993, 1994) and in coupling TRH-R to PLC activation in Xenopus laevis oocytes [de la Peña et al. (1995) but see Quick et al. (1994) and Stehno-Bittel et al. (1995)]. The route or routes linking TRH-R to pituitary cells MAPK activation are not completely defined. Furthermore, the nature of the transducer or transducers implicated in this effect is not known (Ohmichi et al., 1994). To gain further insights into these issues, we used COS-7 cells as a model, in which both TRH-R and an epitope-tagged MAPK are exogenous elements introduced by transfection. Our results demonstrate a clear stimulation of the kinase by TRH in a time- and concentration-dependent manner. This contrasts with the lack of MAPK activation reported previously using Rat-1 cells transfected with the TRH-R (Lee et al., 1995). Furthermore, it allows the performance of studies of the molecular components and characteristics of the coupling cascade started in TRH-R with a cellular model widely used to define these entities.

The pioneering studies in pituitary cells (Ohmichi et al., 1994) indicated that a PKC-dependent pathway coexisted with a second route to MAPK activation independent of PKC activity, possibly involving tyrosine phosphorylation of Shc proteins, and in which activated Raf-1 and p21^ras presumably played a role. That a PKC-dependent pathway also partially links TRH-R to MAPK activation in COS-7 cells is indicated by our results obtained after inhibition of PKC. Initially, we used a procedure for depletion of conventional PKC through prolonged treatment with a high concentration of PMA. Our data, although consistent with a participation of PKC in TRH effects, were complicated by possible secondary effects of this treatment that resulted in high basal MAPK activities to a varying extent even before the addition of an agonist. Thus, it is possible that alterations of cellular components during the prolonged exposure to PMA or maintenance of a small fraction of highly active PKC after the long PMA treatment contributed to the observed results. As an alternative, we used the nontoxic PKC-specific inhibitor GF109203X under conditions previously shown to effectively block phosphorylation of endogenous PKC substrates as well as biological responses induced by phorbol esters in COS-7 cells (Crespo et al., 1994b; Coso et al., 1995). GF109203X pretreatment inhibited MAPK activation without affecting basal activities. However, TRH-induced activation was only partially inhibited, indicating that the effect of the hormone involves both PKC-dependent and -independent mechanisms. It is important to note that the partial effect of GF109203X is not due to a limited effect of the inhibitor on PKC enzymatic activity, because the exclusively PKC-dependent action exerted by PMA was almost completely abolished by GF109203X.

Depending on the receptor, MAPK activation in COS-7 cells may be mediated by PTX-sensitive or -insensitive G proteins and be either p21^ras or PKC dependent, respectively. Thus, MAPK activation via _2A-adrenergic, m2 muscarinic acetylcholine, D₂ dopamine, A1 adenosine, and endogenous lysophosphatidic acid receptors is mediated by subunits derived from PTX-sensitive G proteins in a p21^ras-dependent manner (Luttrell et al., 1996, and references therein). In contrast, PTX-insensitive signaling to MAPK by G_q-coupled _1B-adrenergic and m1 muscarinic receptors has been reported to be insensitive to expression of a dominant negative mutant of p21^ras and of a G-sequestrant -adrenergic receptor kinase peptide (Faure et al., 1994; Koch et al., 1994; Hawes et al., 1995). This has been interpreted as indicating that MAPK activation by G_q-coupled receptors is G_q mediated, PKC dependent, p21^ras independent, and not dissociable from the ability of these receptors to activate PLC (Hawes et al., 1995). However, only a partial dependence on PKC activity has been reported by other authors for m1-R (Crespo et al., 1994b), and some inhibition of m1-R-activating effects by -sequestering t subunits has been observed in COS-7-transfected cells (Crespo et al., 1994a). Our results indicate that although totally insensitive to PTX treatment, the TRH-induced increases in MAPK activity are in part transduced by a -mediated and p21^ras-dependent pathway. Thus, (1) only a fraction of the TRH-induced activation was blocked by a PKC inhibitor that abolished PKC-dependent activations triggered by treatment with PMA; (2) activation was inhibited partially by overexpression of t subunits, and (3) expression of the dominant negative p21^N17ras form of p21^ras inhibited TRH-induced activation of MAPK but not the PKC-dependent activation induced by PMA. As for the PKC inhibitor, the reduction in TRH-induced activation by p21^N17ras was only partial in the same cells in which the totally Ras-dependent activation initiated by the m2-R was almost abolished by the Ras competitor. The additivity of p21^N17ras and PKC inhibitor GF109203X effects indicates that the p21^ras-mediated route differs from the PKC-dependent pathway initiated in the TRH-R. This is supported further by our results showing an additive inhibition caused by the PKC inhibitor and a subunit scavenger (t) but not by p21^N17ras and t, which presumably act at different levels of the same Ras-dependent route. These results demonstrate that not only can receptors coupled to PTX-sensitive G_i or G_o use dimers to signal for MAPK activation but also a similar pathway can be used, at least in part, for MAPK activity increases transduced exclusively by G proteins insensitive to PTX and CTX. Interestingly, a p21^ras-dependent route regulated by PTX-insensitive G protein or proteins has been reported for thrombin and angiotensin II receptors in fibroblasts and cardiac myocytes, respectively (Chen et al., 1996; Sadoshima et al., 1996).

Although insensitive to PTX and CTX, the identity of the G protein or proteins coupling TRH-R to MAPK activation is not known. Our efforts to advance in this direction were greatly hampered by the inability to mimic the TRH effects with mutationally activated forms or to potentiate the hormonal effects with wild-type forms of specific  subunits. Failure to detect any effect of dimer sequestrants on G_q-mediated responses has been taken as an indication that only subunits of PTX-sensitive G_i or G_o, and not those of G_q/11, are able to transduce signals to MAPK (Crespo et al., 1994b; Russell et al., 1994). However, this interpretation is challenged by our results showing that a fraction of the toxin-insensitive effect in response to TRH is mediated by subunits. This would be consistent with the reported tyrosine phosphorylation of Shc leading to p21^ras and Raf-1 activation in pituitary GH₃ cells, presumably involving G_q/11, although the toxin sensitivity of this TRH-induced effect was not investigated (Ohmichi et al., 1994). Nevertheless, the possibility cannot be completely excluded that subunits released from a TRH-R-coupled toxin-insensitive G protein other than G_q/11 also contribute to TRH effects on MAPK.

It is known that introduction in COS cells of the constitutively active from of G_q (q- QL) persistently activates PLC, causing a sustained high steady state level of inositol phosphates to levels similar to those reached by stimulation of G_q-coupled receptors such as the m1-R (Qian et al., 1993). However, a similar activation of MAPK is not observed even though clearly PKC-dependent effects are obtained in response to PMA or to stimulation of TRH-R and m1-R. A plausible hypothesis to explain this apparent paradox is that long term stimulation of one or several cell components causes their inactivation at some time after the appearance of the permanently active q-QL. Because prolonged activation of PKC by PMA leads to down-regulation of the enzyme, it remained possible that a similar mechanism due to sustained elevation of diacylglycerol participates in prevention of MAPK activation by q-QL. This would be compatible with the q-QL-induced inhibition of the PKC-mediated activation promoted by PMA. However, our results provide a more straightforward interpretation of the data: the inhibitory effect of q-QL expression is exerted at a step downstream of that at which PKC-dependent and -independent routes to MAPK converge (e.g., at the level of Raf-1 or MAPK). This is supported by the similar inhibition of the PKC-independent activations in response to stimulation of EGF or m2 muscarinic receptors. Further work is necessary to localize the specific step or steps of the cascade inhibited in the presence of q-QL.

In summary, as shown previously in GH₃ anterior pituitary cells (Ohmichi et al., 1994), TRH-Rs activate MAPK in transfected COS-7 cells via a dual pathway only partially dependent on PKC activity. subunits of a PTX- and CTX-insensitive G protein are involved in the PKC-independent fraction of the signaling route to MAPK initiated in the TRH-R. Thus, it is tempting to hypothesize that a similar -dependent mechanism is the cause of the PKC-independent fraction of MAPK activation involving tyrosine phosphorylation in pituitary cells. The identity of the G protein or proteins coupling TRH-Rs to MAPK activation remains to be established. Finally, although the relevance of TRH-induced MAPK activation in pituitary cell physiology is not known, we verified recently that, as previously shown for other G protein-coupled receptors (Crespo et al., 1994a), prolonged challenge with TRH leads to the formation of proliferation foci in transfected NIH 3T3 cells (P. de la Peña and J. S. Gutkind, unpublished observations). This opens up the possibility that some of the TRH effects can be related to its action as a proliferative factor under certain physiological or pathological conditions.