Differential Activation of Protein Kinase C d and e Gene Expression by Gonadotropin-releasing Hormone in aT3-1 Cells

The effect of gonadotropin-releasing hormone (GnRH) upon protein kinase C (PKC) δ and PKCε gene expression was investigated in the gonadotroph-derived αT3-1 cell line. Stimulation of the cells with a stable analog [d-Trp6]GnRH (GnRH-A) resulted in a rapid elevation of PKCε mRNA levels (1 h), while PKCδ mRNA levels were elevated only after 24 h of incubation. The rapid elevation of PKCε mRNA by GnRH-A was blocked by pretreatment with a GnRH antagonist or actinomycin D. The PKC activator 12-O-tetradecanoylphorbol-13-acetate (TPA), but not the Ca2+ ionophore ionomycin, mimicked the rapid effect of GnRH-A upon PKCε mRNA elevation. Additionally, the rapid stimulatory effect of GnRH-A was blocked by the selective PKC inhibitor GF109203X, by TPA-mediated down-regulation of endogenous PKC, or by Ca2+ removal. Interestingly, serum-starvation (24 h) advanced the stimulation of PKCδ mRNA levels by GnRH-A and the effect could be detected at 1 h of incubation. The rapid effect of GnRH-A upon PKCδ mRNA levels in serum-starved cells was mimicked by TPA, but not by ionomycin, and was abolished by down-regulation of PKC or by Ca2+ removal. Preactivation of αT3-1 cells with GnRH-A for 1 h followed by removal of ligand and serum resulted in elevation of PKCδ mRNA levels after 24 h of incubation. Western blot analysis revealed that GnRH-A and TPA stimulated (within 5 min) the activation and some degradation of PKCδ and PKCε. We conclude that Ca2+ and PKC are involved in GnRH-A elevation of PKCδ and PKCε mRNA levels, with Ca2+ being necessary but not sufficient, while PKC is both necessary and sufficient to mediate the GnRH-A response. A serum factor masks PKCδ but not PKCε mRNA elevation by GnRH-A, and its removal exposes preactivation of PKCδ mRNA by GnRH-A which can be memorized for 24 h. PKCδ and PKCε gene expression evoked by GnRH-A is autoregulated by PKC, and both isotypes might participate in the neurohormone action.

The protein kinase C (PKC) 1 family is a family of serine/ threonine protein kinase isoforms, which play key roles in signal transduction (1)(2)(3). Conventional PKCs (␣, ␤I, ␤II, and ␥) are activated by Ca 2ϩ , diacylglycerol (DAG), and phospholipid such as phosphatidylserine (PS) and are tightly coupled to phosphoinositide turnover (1)(2)(3). Novel PKCs (␦, ⑀, , and ) are Ca 2ϩ -independent but DAG-and PS-activated isoforms. Atypical PKCs ( and /) are Ca 2ϩ -and DAG-independent but PS-activated isoforms and are also stimulated by other lipidderived mediators (1)(2)(3)(4). PKC takes an intermediate position among the novel PKC and atypical PKC isoforms and is a Ca 2ϩand DAG-independent isoform. Whereas relatively much is known about regulation of PKC at the protein level including cofactor requirements, translocation to the membrane, substrate phosphorylation, and degradation (1-9), very little is known about ligand regulation of PKC gene expression (10 -12). Previous work has implicated PKC in gonadotropin-releasing hormone (GnRH) action upon gonadotropin secretion and gonadotropin subunits gene expression in pituitary and ␣T3-1 cells (5,6,(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26). Recently, while examining conventional PKC regulation, we have shown that GnRH-A increases the levels of PKC␤, but not PKC␣, mRNA levels in ␣T3-1 cells, while PKC␥ is not expressed in the cells (12). Since PKC␦ and PKC⑀ of the novel PKC group are major subspecies in the pituitary (26), we decided to investigate the effect of GnRH-A on the mRNA levels of both isotypes in the ␣T3-1 cell line. Here we demonstrate that GnRH-A directs differential autoregulation of PKC␦ and PKC⑀ gene expression, which is dependent upon growth conditions and Ca 2ϩ , and reveals a memory mechanism, which might participate in PKC␦ autoregulation.   (27), and the respective antibodies were obtained from Sigma.

Methods
Cell Culture-␣T3-1 cells were subcultured into 60-mm tissue culture dishes (Sterilin, Hounslow, United Kingdom). Cells were grown in 5 ml of Dulbecco's modified Eagle's medium (DMEM) containing 5% fetal calf serum, 5% horse serum, 100 units/ml penicillin, and 0.1 mg/ml streptomycin. After 3-4 days, when cells were 70 -80% confluent, the cultures were washed three times with fresh DMEM, and stimulants were added in 5 ml of DMEM at the indicated concentrations for the given length of time. For short period incubations (up to 1 h) 10 mM Hepes was added to the medium. When the stimulation period was longer than 9 h, the medium was supplemented with 0.1% bovine serum albumin.
RNA Extraction and Analysis-At the end of the stimulation period, total RNA was isolated from cells by extraction in guanidium thiocyanate containing 8% 2-mercaptoethanol by the LiCl method as described by Cathala et al. (28). For Northern blot analysis, total RNA (15 g) was fractionated on 1.2% denaturing agarose gel and transferred to Gene-Screen membranes (DuPont NEN). Alternatively, RNA samples (8 g) were slot blotted onto GeneScreen using a slot blot manifold (Schleicher & Schü ll). Following baking and prehybridization, the membranes were hybridized overnight with the specific cDNA probes labeled to high specific activity using a random primer labeling kit (Boehringer). Half of each lane was hybridized with a PKC cDNA, and the second half was hybridized with glyceraldehyde-3-phosphate dehydrogenase cDNA as an internal control. Thereafter, filters were washed at high stringency and were autoradiographed at Ϫ70°C. Steady state levels of mRNAs were quantified with densitometric scanning of autoradiograms. The data were corrected for variability in loading by calculation as a ratio to glyceraldehyde-3-phosphate dehydrogenase.
PKC Translocation-Following ligand treatment, cells were washed with ice-cold Tris-buffered saline, pH 7.2, harvested with rubber policemen, and pelleted by short spin (1200 rpm for 5 min at 4°C). Cells were resuspended in 10 mM EGTA, 2 mM EDTA, 20 mM Tris-HCl, pH 7.5, 0.25 M sucrose, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 0.5 mM iodoacetic acid, and lysed by 10 strokes of a 25-gauge syringe. Following removal of nuclei (1200 rpm for 5 min at 4°C), cytosol and membrane fractions were obtained by ultracentrifugation (100,000 ϫ g for 2 h at 4°C). The proteins were separated on 7.5-18% SDS-polyacrylamide gels (ratio of acrylamide to bisacrylamide, 30:0.5) and electrotransferred to nitrocellulose papers in 50 mM glycine, 50 mM Tris-HCl, pH 8.8 (100 V for 2 h at 4°C). The papers were blocked for 60 min in 1% bovine serum albumin and 0.5% Tween 20 in Tris-buffered saline and treated overnight with the respective rabbit anti-PKC antibodies (Sigma). The signals were visualized using horseradish peroxidase-conjugated goat anti-rabbit IgG and the ECL method.
Statistical Analysis-The hybridization signals for PKC subtypes mRNA in each group were normalized to the hybridization signals for the housekeeping gene for glyceraldehyde-3-phosphate dehydrogenase. An arbitrary unit of 1 represents the control values. Statistical compar-isons between control and treatment groups were performed using Student's t test; in the figures, a single asterisk indicates p Ͻ 0.05, a double asterisk indicates p Ͻ 0.01, and a triple asterisk indicates p Ͻ 0.001.

RESULTS
We first studied the cellular redistribution of PKC␦ and PKC⑀ following GnRH-A and TPA stimulation, since it is a criterion for PKC activation by extracellular signals (5)(6)(7)(8). Both GnRH-A and TPA stimulated an increase in the molecular weight of cytosolic PKC␦ within 5 min (Fig. 1), consistent with the size shift reported for PKC␦ phosphorylation by Src (29), and with our own finding that GnRH-A and TPA stimulate protein-tyrosine phosphorylation in ␣T3-1 cells. 2 PKC␦ in the membrane fraction is already of the high molecular weight form and is further elevated by GnRH-A and even more by TPA. In addition, translocation of PKC␦ to the membrane fraction by GnRH-A and TPA is further validated by the appearance of degradation products of 70 and 42 kDa (apparently PKM; Refs. 6 -8) in the membrane fraction in the ligandtreated groups (3-4-fold stimulation by GnRH-A and TPA; Fig.  1). PKC⑀ activation is manifested by translocation to the membrane fraction and the appearance of 50-and 42-kDa bands (apparently PKM) in the ligand-treated groups (2-fold; Fig. 1). Consistent with our previous reports that TPA-mediated downregulation of endogenous PKC in ␣T3-1 cells reduced cellular PKC activity by 90% (12,21,23), prolonged incubation with TPA (100 ng/ml, 24 h) resulted in loss of most of PKC⑀ (60 and 90% of the membrane and soluble enzyme, respectively), and all of the detectable soluble and membrane-bound PKC␦ (Fig. 1).
The regulation of PKC␦ and PKC⑀ mRNA levels was determined by treatment of ␣T3-1 cells with [D-Trp 6 ]GnRH, a stable GnRH analog. Addition of GnRH-A to the cells for 1 h elevated PKC⑀ but not PKC␦ mRNA levels in a dose-related fashion, 2 N. Reiss, D. Harris, and Z. Naor, manuscript in preparation. with maximal response obtained at 10 nM analog ( Fig. 2A). Stimulation of PKC⑀ mRNA levels by GnRH-A was rapid, with a peak at 1 h, declining thereafter to basal levels (Fig. 2B). On the other hand, significant elevation of PKC␦ mRNA levels was detected only after 24 h of incubation with GnRH-A (3-fold, p Ͻ 0.001; Fig. 2B).
The effect of GnRH-A on PKC⑀ mRNA levels was investigated further, since its rapid nature suggested that it more likely represents a physiological response to the neurohormone. Pretreatment of the cells with a potent GnRH antagonist (Fig. 3A) or with actinomycin D (Fig. 3B) abolished the stimulatory effect of GnRH-A upon PKC⑀ mRNA levels, indicating a receptor-mediated effect apparently at the transcriptional level (Fig. 3).
The potential role of PKC and Ca 2ϩ in mediating the GnRH response upon PKC⑀ mRNA levels was investigated since both messengers were implicated in GnRH action upon gonadotropin release and gonadotropin subunits gene expression (5,6,(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)30). Addition of the PKC activator TPA to ␣T3-1 cells for 1 h resulted in elevation of PKC⑀ but not PKC␦ mRNA levels, while the Ca 2ϩ ionophore, ionomycin, had no effect (Fig. 4, A and C, and data not shown). Elevation of PKC⑀ mRNA levels by TPA was rapid, with a peak at 1 h and a return to basal levels (Fig. 4B). The similar time responses elicited by GnRH-A and TPA suggest that PKC is involved in GnRH-A stimulation of PKC⑀ gene expression.
This notion was further supported by inhibition and depletion of PKC. Addition of the selective PKC inhibitor GF 109203X (27,31) to the cells resulted in a dose-related inhibition of the GnRH-A stimulated PKC⑀ mRNA levels with halfmaximal inhibition (IC 50 ) observed at 0.8 M of the drug, in good agreement with IC 50 values of PKC inhibition in cellular systems such as Swiss 3T3 fibroblasts (31) (Fig. 5A). The drug alone (1 M) reduced the basal level by about 50%, suggesting that PKC is also involved in the maintenance of basal PKC⑀ gene expression. We also used down-regulation of endogenous PKC by prolonged incubation with TPA. Pretreatment of the cells with TPA (100 ng/ml, 24 h) reduced cellular PKC activity by 90% as measured by enzymatic activity assay and Western blot analysis ( Fig. 1 and Refs. 12, 21, and 23). The stimulatory effect of GnRH-A and TPA upon PKC⑀ mRNA levels was abolished in the down-regulated cells (Fig. 5B). In addition, we observed no additivity between GnRH-A and TPA upon PKC⑀ mRNA levels (Fig. 6), lending further support to the role of PKC in mediating the GnRH-A effect on PKC⑀ gene expression. The Ca 2ϩ ionophore, ionomycin, had no effect on basal PKC⑀ mRNA levels or on the stimulatory response elicited by GnRH-A or TPA (Fig. 6). On the other hand, transfer of ␣T3-1 cells to Ca 2ϩ free medium, in the presence or absence of EGTA, abolished stimulation of PKC⑀ mRNA levels by GnRH-A (Fig.  7). It therefore seems that Ca 2ϩ is necessary but not sufficient for mediation of the GnRH-A response.
Since GnRH stimulated PKC␦ mRNA levels only after 24 h of incubation in medium without serum (Fig. 2B), it was possible that growth conditions are involved in PKC␦ gene expression.
We therefore examined the role of serum in PKC⑀ and PKC␦ gene expression. Transfer of the cells to medium with low serum (0.5%) for 24 h had no effect on GnRH-A-stimulation of PKC⑀ mRNA levels (Fig. 8). On the other hand, serum starvation advanced the stimulation of PKC␦ mRNA levels by GnRH-A to 1 h of incubation that could not be observed in serum-grown cells (Fig. 8). Time course of PKC␦ mRNA levels in serum-starved cells revealed a rapid effect of GnRH-A at 1 h of incubation with no effect at 24 h, as seen in non-starved cells (Fig. 9A). Similarly, serum starvation exposed a rapid response (peak at 30 min) of TPA on PKC␦ mRNA levels (Fig. 9B), suggesting a role for PKC in mediating PKC␦ gene expression. Indeed, down-regulation of endogenous PKC by prolonged incubation with TPA abolished GnRH-A and TPA stimulation of PKC␦ mRNA levels in serum-starved cells (Fig. 10). Transfer of the cells to Ca 2ϩ free medium, in the presence or absence of EGTA, abolished the rapid stimulation of PKC␦ in serumstarved cells (Fig. 11).
As shown above, when cells are transferred to serum-free medium and exposed to GnRH-A, elevation of PKC␦ mRNA levels is observed at 24 h, but not at 1 h of incubation (Fig. 12,  columns 1-3). On the other hand, when cells are first serumstarved (24 h) and later exposed to GnRH-A, elevation of PKC␦ mRNA levels is observed after 1 but not 24 h of incubation (  12, columns 4 -6). We therefore exposed normal cells to GnRH-A for 1 h, washed the cells several times to remove serum and GnRH-A, and further incubated the cells for 24 h. As seen in Fig. 12 (columns 7 and 8), PKC␦ mRNA levels were elevated at 24 h by pretreatment (1 h) with GnRH-A. Thus, the late effect of GnRH-A on PKC␦ mRNA levels (Fig. 12, column 3) is due to generation of a rapid signal (1 h, column 5), which is "memorized" during the long starvation period (t ½ Х 12 h) required for manifestation of the early signal by means of removal of the inhibitory effect of the serum. DISCUSSION Whereas much has been learned concerning the regulation of PKC and its subspecies at the protein level (1-9), very little is known about ligand regulation of PKC subtypes gene expression (10 -12). Differentiation regulators of the human promyelocytic leukemia cell line (HL-60) such as 1␣,25-dihydroxyvitamin D 3 , retinoic acid, and dimethyl sulfoxide, were shown to increase the expression of PKC␣ and PKC␤ mRNA levels (10,11). Furthermore, transcriptional activation of PKC␣ and PKC␤ expression was reported to result in increased PKC enzymatic activity (10,11,32). Here we demonstrate that GnRH-A, which does not promote growth or differentiation, is capable of activating differential nPKC isoforms gene expression. To the best of our knowledge, this is the first demonstration of a natural ligand stimulation of PKC␦ and PKC⑀ mRNA levels.
Activation of PKC␦ mRNA levels, but not that of PKC⑀, is dependent upon growth conditions, suggesting the presence of a serum factor that is involved in regulation of PKC␦ gene expression, possibly via a serum response element.
The differential activation of PKC␦ and PKC⑀ mRNA levels by GnRH-A suggests that the isoforms might specialize in different functions. PKC␦ is the major subspecies in the 6-dayold rat pituitary and is markedly reduced in the 3-month-old pituitary (26). The opposite is observed for pituitary PKC⑀, which increases with age (26). Therefore, PKC␦ and PKC⑀ might play different roles during pituitary development. It was also shown that while PKC␦ is involved in exocytosis, PKC⑀ participates in feed-back inhibition of phospholipase C activity in rat basophilic RBL-2H3 cells (33). In a recent study, GnRH was shown to translocate PKC⑀ and PKC in ␣T3-1 cells while PKC␦ was not detected (30). We report here that both the PKC␦ and PKC⑀ isoforms are expressed in the ␣T3-1 cells at the mRNA and protein levels and that they are translocated to the membranes and activated in response to GnRH-A, as also validated by the apparent formation of the PKM species (6 -8). The differences between the reports are most likely due to the use of different PKC type-specific antibodies.
While overexpression of PKC␦ resulted in inhibition of growth rate in NIH 3T3 cells, overexpression of PKC⑀ increased growth rate, and the transformed cells (NIH 3T3 or Rat 6 cells) formed tumors in nude mice (27,34). Since GnRH affect differ-entiated responses, it is possible that PKC␦ and PKC⑀ are involved in separate functions such as gonadotropin release and gonadotropin subunit gene expression during the hormone action. Elevation of mRNA of a given PKC isoform by ligands in general and by GnRH-A in particular might be a step in the life cycle of PKCs during hormone action to replenish the enzymes after translocation and degradation as shown in Fig. 1.
The rapid effect (peak at 60 min) of GnRH-A upon PKC⑀ and PKC␦ (in serum-starved cells) mRNA levels might be physiologically relevant, since GnRH is released from the hypothalamus in a pulsatile manner at intervals of 1-2 h according to the species and its half-life is about 2-4 min (35)(36)(37). Thus, prolonged responses such as those observed in Fig. 2 (24 h) are more difficult to interpret, since it is not clear whether multiple pulses of GnRH are capable of eliciting a similar response. Hence, we investigated in more detail the rapid effects of GnRH, which prompted us to identify the second messengers involved in the neurohormone action. Indeed, the PKC activator TPA mimicked the GnRH-A rapid responses and stimulated PKC⑀ mRNA levels in serum-grown cells and PKC␦ mRNA levels in serum-starved cells with a similar time course. Additionally, the stimulatory effect of GnRH-A on PKC␦ and PKC⑀ mRNA levels was abolished in PKC-down-regulated cells or by the use of the selective PKC inhibitor GF 109203x (27,31) (present results and data not shown). We therefore suggest that GnRH-A stimulation of PKC␦ and PKC⑀ gene expression is autoregulated by PKC.
While we show here positive regulation of PKC␦ mRNA levels by GnRH-A and TPA, others have recently reported that PKC␦ mRNA is down-regulated by TPA (4 h) in the mouse B lymphoma cell line A20 (38). The difference between the results might be due to the presence of a soluble destabilizing factor, which specifically accelerates degradation of PKC␦ mRNA in A20 cells (38).
Removal of Ca 2ϩ abolished the effect of GnRH-A upon PKC␦ and PKC⑀ mRNA levels, but Ca 2ϩ ionophore had no stimulatory effect. We therefore conclude that Ca 2ϩ is necessary but not sufficient, while PKC is both necessary and sufficient to mediate the GnRH response. Furthermore, since removal of extracellular Ca 2ϩ per se is not sufficient to block Ca 2ϩ mobi-  11. Effect of Ca 2؉ removal upon GnRH-A-induced PKC␦ mRNA levels in serum-starved cells. ␣T3-1 cells were preincubated with 0.5% serum for 24 h. Cells were then transferred to DMEM (control)), to Ca 2ϩ -free DMEM (Ca 2ϩ free) or to Ca 2ϩ -free DMEM ϩ 250 M EGTA (EGTA) for 10 min. Cells were then incubated with (striped bars) or without (empty bars) GnRH-A (10 nM) for 60 min in the respective medium as indicated. PKC␦ mRNA levels were analyzed as described under "Methods" (n ϭ 6). ** p Ͻ 0.01. lization in pituitary cells (39), the data suggest that GnRH-Ainduced PKC␦ and PKC⑀ gene expression is mainly mediated by Ca 2ϩ influx, apparently via L-type voltage-sensitive channels (39). The data also suggest that a Ca 2ϩ -dependent PKC isoform might be involved in GnRH-A action. Since PKC␦ and PKC⑀ are differentially regulated by GnRH-A, it is unlikely that one regulates the expression of the other during the neurohormone action. Furthermore, it is unlikely that PKC⑀ is involved in the process since its membrane-bound form (the active form) was reduced only by 60% by down-regulation, whereas the effect of GnRH-A was abolished. Further studies are required to identify the PKC isoforms and the site of Ca 2ϩ action in the GnRH-A response.
In addition to mediation by Ca 2ϩ and PKC, elevation of PKC␦ but not PKC⑀ mRNA levels by GnRH-A and TPA required removal of a serum factor. Therefore, although PKC␦ and PKC⑀ gene expression share some common mechanisms, which are mediated by Ca 2ϩ and PKC, they differ in sensitivity to the serum factor. Changes in the concentrations of the serum factor under physiological conditions might therefore enable preferential activation by GnRH of the two isotypes. Moreover, it was possible to first stimulate the cells with GnRH-A, and remove the hormone and serum for 24 h, at the end of which elevation of PKC␦ mRNA levels by GnRH-A was detected. The observation is in contrast to previous observations, in which GnRH-stimulated LH release in perifused pituitary cells was terminated immediately after removal of the neurohormone (Ref. 40 and data not shown). Since PKC is the main mediator of GnRH actions (5,6,(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)30), it is likely that the half-life of the phosphoproteins involved in exocytosis is very short (t ½ Х 8 min; Ref. 40), while those mediating the neurohormone effect on PKC␦ gene expression is relatively long (t ½ Х 12 h). The presence of the serum factor does not block the formation of downstream effectors involved in PKC␦ gene expression, but only masks the effectors activity. Since growth conditions also affected TPA stimulation of PKC␦ gene expression, it seems that the site of action of the serum factor is downstream to PKC activation. Since both Ca 2ϩ and PKC participate in GnRHstimulated PKC␦ and PKC⑀ gene expression, it is likely that transcriptional regulation of both isotypes by GnRH-A involves similar transcription factors but different coactivators and composite response elements (41). Future analysis of the gene structure of both isotypes will reveal the different response elements involved in ligand regulation of PKC␦ and PKC⑀ gene expression.
The present report demonstrates differential activation of PKC␦ and PKC⑀ mRNA levels by GnRH-A that is dependent upon growth conditions and Ca 2ϩ influx, and is autoregulated by PKC. Since both isotypes are shown here to be translocated by GnRH from the cytosol to the membrane (an index of PKC activation; Refs. 1-3 and 5-8), PKC␦ and PKC⑀ are therefore likely candidates to participate in GnRH action, the first key hormone of the reproductive cycle.