Ethanol Stimulates cAMP-Responsive Element (CRE)-Mediated Transcription via CRE-Binding Protein and cAMP-Dependent Protein Kinase

Abstract

Alcoholism is characterized by tolerance, dependence, and unrestrained craving for alcohol. Adaptive responses, including changes in gene expression in neurons, are thought to account for some of these complex behavioral abnormalities. We have shown in the NG108-15 neuroblastoma × glioma hybrid cell line that ethanol increases cellular cAMP levels via activation of adenosine A2receptors, leading to phosphorylation of the cAMP response element-binding protein (CREB). However, phosphorylation of CREB is not sufficient to activate cAMP response element (CRE)-mediated gene expression. Here we investigate whether ethanol increases CRE-mediated gene expression via endogenous CREB using a CRE-regulated luciferase reporter construct, transfected into NG108-15 cells. We find increased luciferase activity as a function of time of exposure to ethanol. Coexpression of a dominant-negative CREB construct blocked ethanol-stimulated CRE-luciferase expression, further suggesting that CREB is required for this response. We also determined whether ethanol-induced increases in gene expression are mediated by ethanol-induced increases in extracellular adenosine. We found that CRE-mediated gene expression induced by ethanol occurs in two phases: an early phase (4 h), in which adenosine receptor blockade prevents ethanol-induced gene expression, and a later phase (14 h), which is not blocked by an adenosine receptor antagonist. In both phases, inhibition of cAMP-dependent protein kinase A (PKA) activity prevented ethanol-induced CRE-mediated luciferase expression. Our data suggest that ethanol induces cAMP-dependent gene expression regulated by CREB and PKA and that this signaling pathway may mediate some of the addictive behaviors underlying alcoholism.

Numerous studies have demonstrated that multiple intracellular signaling pathways regulate gene transcription. cAMP-dependent protein kinase A (PKA) increases gene expression through the cAMP response element-binding protein (CREB) (Sassone-Corsi, 1995; Montminy, 1997), which is an important component in long-term changes in synaptic plasticity and memory (Bailey et al., 1996; Silva et al., 1998). CREB phosphorylation on Ser-133 promotes activation of CREB and transcription of genes with an upstream cAMP response element (CRE) in cell lines and in primary neuronal cultures (Impey et al., 1998; Tao et al., 1998). CREB phosphorylation is regulated by PKA and by other protein kinases such as Ca2+/calmodulin-dependent kinase (CaMK) (Matthews et al., 1994), and protein kinase C (PKC) (Muthusamy and Leiden, 1998; Roberson et al., 1999). Thus, CREB is a target for several different signaling pathways (Matthews et al., 1994;Impey et al., 1998; Muthusamy and Leiden, 1998; Roberson et al., 1999). Furthermore, cAMP signaling and phosphorylation of CREB is implicated in alcohol-related behaviors (Moore et al., 1998; Thiele et al., 2000;Pandey et al., 2001; Wand et al., 2001).

Acute and chronic exposure to ethanol alters the activity of several signal transduction systems (Diamond and Gordon, 1997). In NG108-15 neuroblastoma × glioma cells, acute exposure to ethanol increases cAMP production, which is blocked by adenosine receptor antagonists (Sapru et al., 1994). Our laboratory has also shown that chronic exposure to ethanol causes sustained translocation of the catalytic subunit of PKA (Cα) to the nucleus in NG108-15 cells (Dohrman et al., 1996); Cα remains in the nucleus as long as ethanol is present. Ethanol also causes a prolonged increase in CREB phosphorylation in NG108-15 cells (Constantinescu et al., 1999). However, CREB phosphorylation is not sufficient to activate gene expression (Montminy, 1997; Cardinaux et al., 2000). Therefore, we investigated directly whether ethanol increases gene expression. We show here that ethanol exposure causes a striking increase in CRE- and CREB-mediated gene expression in cells transfected with luciferase reporter genes and that PKA and CREB are required for this response.

Materials and Methods

Cell Culture.

NG108-15 neuroblastoma × glioma hybrid cells were seeded at a density of 2 × 106cells/ml in T175 flasks and maintained for 48 h (Dohrman et al., 1996). The cells were then plated at a density of 5 × 104/ml in 12-well plates in defined medium and incubated for a further 24 h as described previously (Constantinescu et al., 1999). The ethanol concentration used was 200 mM unless otherwise indicated. As in our previous studies, we chose to use this concentration of ethanol to minimize the time required for the experiments and the use of expensive media components. We have previously shown that there is no effect on cell division or death (Gordon et al., 1986), or on morphology or localization of Golgi markers (Gordon et al., 2001) at this concentration, and the effects of ethanol on PKA signal transduction are reversible within 12 to 16 h after ethanol is withdrawn (Dohrman et al., 1996).

Plasmids.

The constructs pRcRSVCREBM1 (dominant negative CREB) and MT-REVAB (dominant negative PKA-RIα) were kindly provided by Dr. M. E. Greenberg (Tao et al., 1998) and Dr. G. S. McKnight (Clegg et al., 1987), respectively. pFC-CRE-luciferase was purchased from Stratagene (La Jolla, CA); pCMV-β-galactosidase was purchased from Qiagen (Hilden, Germany). No luciferase activity was observed when cells were transfected with the pFC plasmid alone, nor was there any effect on basal- and ethanol-stimulated luciferase activity when cells were cotransfected with the empty vectors of the dominant-negative plasmids.

Transfection Procedures and Reporter Assays.

NG108-15 cells were transfected in defined media with Effectene (Qiagen) as described by the manufacturer. Media were changed 24 h after transfection. To determine the effects of ethanol on CRE-mediated gene expression or GAL4-CREB activation, the cells were then incubated in 200 mM ethanol for various periods of time. For experiments with Rp-cAMPS, the cells were preincubated with 20 μM Rp-cAMPS (BioLog Life Science Institute, La Jolla, CA) for 2 h. The other inhibitors, 10 μM BW A1434U (a gift from GlaxoSmithKline, Uxbridge, Middlesex, UK), 10 μM H-89, 100 nM bisindolylmaleimide I (GF), and 5 μM KN-62 (all purchased from Calbiochem, San Diego, CA) were added 30 min before ethanol exposure; all inhibitors were present during the ethanol incubation. Cell extracts were prepared and luciferase was measured with a commercial assay system (Promega, Madison, WI), using a Rosys Anthos Lucy2 microplate luminometer (Anthos Labtec Instruments, Salzburg, Austria). The results are expressed as relative luciferase activity or as percent increase over control. Data represent three or more separate experiments with triplicate samples in each experiment. Luciferase activities were normalized for transfection efficiency determined in cells transfected in parallel with pCMV-β-galactosidase. β-galactosidase activity was measured using a kit from Stratagene (La Jolla, CA).

Results

Ethanol Stimulates CRE-Mediated Luciferase Expression.

Our laboratory has shown that ethanol increases cAMP levels (Sapru et al., 1994), translocation of PKA to the nucleus (Dohrman et al., 1996), and phosphorylation of CREB (Constantinescu et al., 1999). However, because CREB phosphorylation is not sufficient to activate CRE-mediated gene expression (Montminy, 1997; Cardinaux et al., 2000), we asked whether ethanol increases CRE-mediated gene expression. NG108-15 cells were transiently transfected with a CRE-luciferase reporter construct, and luciferase activity was measured at various times (Fig.1). An increase in luciferase activity was first apparent after a 4-h exposure to 200 mM ethanol (29 ± 6% increase). After 14 h of exposure, luciferase activity increased to 79 ± 4% above control. An increase in luciferase activity (86 ± 6%, n = 4) was also observed after exposure to 100 mM ethanol for 14 h.

Figure 1
View larger version:
Figure 1

Ethanol-induced CRE-mediated luciferase activity as a function of time. NG108-15 cells were transfected with a CRE-luciferase construct (150 ng). Luciferase expression was measured in the presence (▪) or absence (■) of 200 mM ethanol at the indicated times. The results are expressed as relative luciferase activity. Data (±S.E.M.) are the average of three or more separate experiments, each carried out in triplicate. ★, p < 0.01; ★★,p < 0.05

The Role of cAMP and PKA in Ethanol-Induced CRE-Mediated Gene Expression.

We have previously shown that acute exposure of NG108-15 cells to ethanol inhibits adenosine uptake, thereby increasing extracellular adenosine, which activates adenosine A2 receptors and increases intracellular cAMP levels (Nagy et al., 1991). The adenosine receptor antagonist, BW A1434U, prevents this ethanol-induced increase in cAMP (Sapru et al., 1994). It seemed possible, therefore, that the ethanol-induced CRE-mediated gene expression observed in Fig. 1 was due to adenosine receptor-dependent increases in cAMP levels. To test this hypothesis, we coincubated cells transfected with the CRE-luciferase plasmid with ethanol and the nonselective adenosine receptor antagonist BW A1434U, and then measured luciferase activity (Fig.2). BW A1434U blocked the ethanol-induced increase in luciferase activity at 4 h but had no effect on the increase measured at 14 h. Therefore, sensitivity to adenosine receptor blockade appears to distinguish two phases of ethanol-dependent increases in luciferase activity: an early phase, which requires the adenosine A2 receptor (no A1 receptors are present in this cell line; A. S. Gordon and I. Diamond, unpublished observation), and a later phase which is independent of adenosinergic mechanisms.

Figure 2
View larger version:
Figure 2

Regulation of ethanol-induced CRE-luciferase activation. NG108-15 cells were transfected with a CRE-luciferase construct. BW A1434U (10 μM) was added 30 min before and Rp-cAMPS (20 μM) 2 h before 200 mM ethanol exposure. Luciferase activity was measured after a 4-h exposure to ethanol (early phase) and after a 14-h exposure to ethanol (late phase) in the presence of the inhibitors. The results are expressed as percent increase over control. Data (±S.E.M.) are the average of three or more separate experiments. ★,p < 0.01

We next determined whether PKA plays a role in ethanol-induced increases in CRE-mediated luciferase expression. Rp-cAMPS, a selective PKA inhibitor, completely inhibited both the early and late phases of increased luciferase activity (Fig. 2). Because the increase in CRE-mediated luciferase expression in the early phase of ethanol exposure appears to be adenosine A2R-dependent (4 h; Fig.2), we focused our attention on the mechanism underlying chronic exposure to ethanol (14 h). We incubated CRE-luciferase-transfected cells with H-89, another PKA inhibitor. H-89 also inhibited ethanol-induced increases in luciferase expression at 14 h, confirming that PKA activity appears to be required for ethanol-dependent increases in CRE-mediated luciferase activity (Fig.3). NG108-15 cells were also cotransfected with the CRE-luciferase construct and a plasmid containing cDNA for the dominant negative form of PKA-RIα (DN-RIα) (Clegg et al., 1987). This construct contains mutations at the two cAMP binding sites in the RIα subunit of PKA, one at position 200 in site A and two at amino acid positions 324 and 332 in site B, which interfere with PKA activation by cAMP. Expression of the PKA-RIα construct substantially inhibited the ethanol-stimulated increase in luciferase activity (Fig. 3).

Figure 3
View larger version:
Figure 3

Ethanol stimulation of CRE-luciferase activity requires PKA activity. NG108-15 cells were transfected with a CRE-luciferase construct. Luciferase activity was measured after 14 h of exposure to ethanol in cells that were pretreated for 30 min in the absence or presence of the PKA inhibitor H-89 (10 μM). In a second set of experiments, cells were cotransfected with a dominant negative PKA-RIα construct (DN-RIα; 5 ng), together with the CRE-luciferase construct (150 ng). These cells were then incubated in the presence or absence of 200 mM ethanol for 14 h, and luciferase activity was measured. In the third set of experiments, cells were pretreated for 30 min with the CaMK inhibitor KN-62 (5 μM), or the PKC inhibitor bisindolymaleimide I (GF) (100 nM). These cells were then further incubated in the presence or absence of 200 mM ethanol for 14 h, and luciferase activity was measured. The results are expressed as percent increase after 14 h exposure to ethanol over control cells not treated with ethanol. Data (±S.E.M.) are the average of three or more separate experiments. ★, p < 0.01

As discussed above, PKC and CaMK can regulate CREB phosphorylation and activity. Therefore, we used inhibitors of these protein kinases to determine their role in ethanol-stimulated CRE expression. GF, a PKC inhibitor, did not inhibit CRE-mediated increases in luciferase activity in ethanol-treated cells (Fig. 3). In addition, the PKC inhibitor calphostin had no significant effect on ethanol-induced luciferase activity (59 ± 14%, n = 3, compared with 79 ± 2, n = 3 for ethanol alone). These data suggest that PKC is not involved in CRE-mediated gene expression induced by ethanol. KN-62, a CaMK inhibitor, attenuated luciferase activity to a lesser extent than PKA inhibitors (Fig. 3). The latter result is not easily interpretable, however, because KN-62 also stimulated basal luciferase activity by 55% (data not shown). Taken together, our data suggest that ethanol-induced increases in CRE-mediated luciferase expression require PKA and not PKC activation, although it is not certain whether CaMK participates in this process under the conditions of our experiments.

CREB phosphorylation is required for ethanol-induced increases in CRE-mediated transcription.

Phosphorylation of CREB at Ser-133 activates CREB and stimulates CRE-dependent transcription. Recent studies from our laboratory have shown that ethanol causes striking PKA-dependent increases in CREB phosphorylation in NG108-15 cells with a peak at 3 h followed by a sustained increase in phospho-CREB even after 24 h of ethanol exposure (Constantinescu et al., 1999). Here, using the CRE-luciferase plasmid, we show a moderate induction in luciferase activity after 4 h exposure to ethanol, followed by a further increase that peaks at 14 h of ethanol treatment (Fig. 1). To determine whether CREB phosphorylation is required for ethanol-induced increases in gene expression, the CREBM1 plasmid was cotransfected into NG108-15 cells along with the CRE-luciferase construct. CREBM1 codes for a mutant CREB in which Ser-133 is converted to an alanine; thus, CREBM1 can still bind to the CRE but cannot be phosphorylated or activated (Tao et al., 1998). Cotransfection and overexpression of CREBM1 reduced luciferase activity in untreated control cells and completely prevented ethanol-induced increases in luciferase activity (Fig. 4). These data, together with the data in Figs. 2 and 3, suggest that PKA-mediated phosphorylation and activation of CREB are required for ethanol-induced increases in luciferase activity.

Figure 4
View larger version:
Figure 4

Ethanol-induced CRE-mediated gene expression is regulated by CREB activation. NG108-15 cells were transfected with CRE-luciferase and/or in the dominant negative CREB (CREB M1; 50 ng) construct. Luciferase activity was measured after 14 h in the presence or absence of ethanol as indicated. The results are expressed as relative luciferase activity. Data (±S.E.M.) are the average of three or more separate experiments. ★, p < 0.01

Discussion

The major finding in this study is that ethanol induces an increase in gene expression via CREB and PKA. This increase in gene expression requires both PKA and CREB phosphorylation. Although we had previously shown that exposure to ethanol resulted in phosphorylation of CREB in NG108-15 cells, there is accumulating evidence that CREB phosphorylation is not sufficient to regulate gene expression under the control of CREs; activation of the coactivator CREB-binding protein (CBP) and other downstream elements is also required for increases in CRE-mediated gene expression (Montminy, 1997; Cardinaux et al., 2000).Thibault et al. (2000) have reported increases in genes the expression of which is known to be cAMP-dependent. However, ethanol activates many different signal transduction pathways in addition to PKA (Diamond and Gordon, 1997), and most genes have regulatory elements activated or inhibited by all of these pathways. Therefore, the experiments presented here are the first demonstration that ethanol directly activates PKA and CREB-mediated expression of a gene under control of a CRE.

Using a CRE-luciferase expression system in NG108-15 cells, we also find that ethanol activates gene expression in two distinct phases with different mechanisms. An early phase (4 h) is distinguished by a requirement for adenosine A2 receptor activation, but the late phase (14 h) does not require adenosine signaling (Fig.2). PKA antagonists abolish both phases of ethanol-stimulated luciferase expression, indicating that PKA activation is required for both phases of ethanol-induced CRE-mediated gene expression (Figs. 2and 3).

In NG108-15 cells, acute exposure to ethanol inhibits adenosine uptake via an ethanol-sensitive nucleoside transporter, causing an increase in extracellular adenosine, which activates adenosine A2 receptors (Sapru et al., 1994; Diamond and Gordon, 1997). This, in turn, leads to an increase in cAMP production (Gordon et al., 1986; Nagy et al., 1991; Sapru et al., 1994; Diamond and Gordon, 1997), translocation of the catalytic subunit of PKA (Cα) into the nucleus (Dohrman et al., 1996), and phosphorylation of CREB (Constantinescu et al., 1999). Previous studies from our laboratory have shown that an adenosine receptor antagonist blocks acute ethanol-induced increases in cAMP levels (Sapru et al., 1994) and CREB phosphorylation (Constantinescu et al., 1999). As expected, therefore, adenosine A2 receptor blockade prevented early increases in ethanol-stimulated luciferase activity (Fig. 2). By contrast, however, an adenosine receptor antagonist did not alter the second phase of CRE-luciferase activation induced by chronic exposure to ethanol (Fig. 2). This is consistent with our data showing that sustained ethanol-induced translocation of PKA into the nucleus at 24 h is not blocked by an adenosine receptor antagonist (Dohrman et al., 1996). Apparently, the later phases of ethanol-stimulated PKA translocation and CRE-mediated gene expression involve molecular pathways that do not depend on adenosine A2receptors.

We show here that PKA activity is required for both the early and late phases of ethanol-stimulated CRE-mediated luciferase activity. The early phase was inhibited by the selective PKA inhibitor Rp-cAMPS, as was the late phase. In addition, the PKA inhibitor H-89, or overexpression of a dominant negative PKA-RIα construct (Fig. 3), inhibited the late phase of ethanol-stimulated CRE-mediated gene expression. These data are consistent with our findings that chronic ethanol exposure induces translocation of the Cα subunit of PKA to the nucleus (Dohrman et al., 1996). The PKC inhibitors GF and calphostin did not alter CRE-mediated transcription. The CaMK inhibitor KN-62 appeared to inhibit ethanol-stimulated CRE expression, but to a lesser extent than the PKA inhibitors. However, KN-62 stimulated basal luciferase activity by about 55% (data not shown), making it difficult to determine whether CaMK plays a role in ethanol-stimulated luciferase expression. Impey et al. (1998) found that CaMK does not contribute to stimulation of CREB-dependent transcription in PC12 cells and hippocampal neurons. Therefore, we favor the interpretation that the moderate inhibition of ethanol-stimulated luciferase expression by KN-62 in our system may be due to KN-62 stimulation of luciferase expression in the absence of ethanol.

We have previously shown that ethanol induces CREB phosphorylation in NG108-15 cells with a peak at 3 h, followed by a sustained elevation of phospho-CREB even after 24 h of ethanol treatment (Constantinescu et al., 1999). In this study, we show that CREB phosphorylation is essential for induction of CRE-mediated gene expression by ethanol by using a dominant negative CREB construct with a mutation at Ser-133 (Fig. 4). Taken together, our data suggest that PKA phosphorylation and activation of CREB are required for ethanol-induced increases in CRE-dependent luciferase expression.

In a recent review, Nestler (2001) suggested that cAMP signaling and CREB phosphorylation may be a compensatory response to addicting drugs. PKA-dependent mechanisms have also been implicated in learning and memory (Bailey et al., 1996; Martin and Kandel, 1996; Silva et al., 1998; Chain et al., 1999). Furthermore, expression profiling of neural cells suggests that many ethanol-regulated genes are also regulated by cAMP (Thibault et al., 2000). All of these findings are consistent with studies that implicate cAMP signaling in alcohol-drinking behavior (Diamond and Gordon, 1997; Moore et al., 1998; Thiele et al., 2000;Pandey et al., 2001; Wand et al., 2001) as well as other studies documenting ethanol-induced changes in CREB phosphorylation in rat striatum and cerebellum (Yang et al., 1998a,b). Our results using luciferase reporter constructs provide the first direct evidence for ethanol-dependent increases in gene expression that require PKA as well as CREB phosphorylation and activation. The molecular changes in the PKA signal transduction pathway, and subsequent changes in gene expression observed in our studies and those of others (Diamond and Gordon, 1997; Thibault et al., 2000), may mediate some of the complex behaviors that underlie alcoholism and addiction.

Acknowledgments

We thank Dr. Anastasia Constantinescu for many helpful discussions concerning the experiments presented here and for a critical reading of the manuscript. We also thank Drs. Jennifer Whistler, Robert O. Messing, and Dorit Ron for comments on the manuscript.

Footnotes

  • This study was supported in part by grants from the State of California for medical research on alcohol and substance abuse through the University of California, San Francisco, and the National Institutes of Health Grant R01 AA10039.

  • Abbreviations:
    PKA
    protein kinase A
    CRE
    cAMP response element
    CREB
    cAMP response element-binding protein
    catalytic subunit of PKA
    CaMK
    Ca2+/calmodulin-dependent kinase
    PKC
    protein kinase C
    GF
    bisindolylmaleimide I
    • Received October 5, 2001.
    • Accepted December 7, 2001.

References

« Previous | Next Article »Table of Contents