Nitric oxide and cGMP regulate gene expression ... - The FASEB Journal

Nitric oxide and cGMP regulate gene expression in neuronal and glial cells by activating type II cGMP-dependent protein kinase TANIMA GUDI,* GREGORY K.-P. HONG,* ARIE B. VAANDRAGER,‡ SUZANNE M. LOHMANN,§ AND RENATE B. PILZ*,1 *University of California, San Diego, and Cancer Center, La Jolla, California 92093-0652, USA; ‡ Erasmus University, 3000 DR Rotterdam, The Netherlands; and §Institut fu¨r Klinische Biochemie und Pathobiochemie, Medizinische Universita¨tsklinik, 97080 Wu¨rzburg, Germany ABSTRACT Nitric oxide (NO) and cGMP have been implicated in many neuronal functions, including regulation of gene expression, but little is known about the downstream targets of NO/cGMP in the nervous system. We found that type II cGMP-dependent protein kinase (G-kinase), which is widely expressed in the brain, mediated NO- and cGMPinduced activation of the fos promoter in cells of neuronal and glial origin; the enzyme was ineffective in regulating gene expression in fibroblast-like cells. The effect of G-kinase II on gene expression did not require calcium uptake but was synergistically enhanced by calcium. G-kinase II was membrane associated and did not translocate to the nucleus; however, a soluble G-kinase II mutant translocated to the nucleus and regulated gene expression in fibroblastlike cells. Soluble G-kinase I also regulates fos promoter activity, but membrane targeting of G-kinase I prevented the enzyme from translocating to the nucleus and regulating transcription in multiple cell types, including glioma cells; this suggests that cell type-specific factor(s) that mediate the transcriptional effects of extranuclear G-kinase II are not regulated by G-kinase I. Our results suggest that G-kinase I and II control gene expression by different mechanisms and that NO effects on neuronal plasticity may involve G-kinase II regulation of gene expression.—Gudi, T., Hong, G. K.-P., Vaandrager, A. B., Lohmann, S. M., Pilz, R. B. Nitric oxide and cGMP regulate gene expression in neuronal and glial cells by activating type II cGMP-dependent protein kinase. FASEB J. 13, 2143–2152 (1999)

Key Words: NO/cGMP signaling z nervous system z fos promoter

In the central nervous system, the nitric oxide (NO)/cGMP signal transduction pathway has been implicated in a wide range of neuronal functions, including modulation of neurotransmitter release and uptake, neuronal differentiation and survival, 0892-6638/99/0013-2143/$02.25 © FASEB

and regulation of synaptic plasticity involved in learning and memory (1–7). Although there are clearly functions of NO in the brain that are independent of cGMP, many physiological effects of NO in the nervous system are attributable to activation of soluble guanylate cyclase and increased intracellular cGMP (2, 4, 8). Compared with other tissues, little is known about the actions of cGMP in the brain. Intracellular targets of cGMP include cGMP-dependent protein kinases (G-kinases), cGMP-regulated phosphodiesterases, and cGMP-gated ion channels, all of which are found in different neuronal tissues and may mediate cGMP effects (9 –12). There are two mammalian G-kinases that are structurally similar and share substrate specificity in vitro, but the enzymes differ in their subcellular localization, tissue distribution, and function (13–15). Type II G-kinase is membrane-bound via an amino-terminal myristoylation site, whereas type I G-kinase is largely cytoplasmic (13, 16). G-kinase II is widely expressed in brain and is also found in kidney and intestinal epithelial cells; G-kinase I is found in cerebellar Purkinje cells, certain sensory neurons, platelets, smooth muscle cells, and endothelial cells (1, 12, 13, 17, 18). Type I G-kinase exists in an a- and b-isoform, which differ only in their first 100 amino acids and are splice products of a single gene (19). The effects of NO/cGMP on neuronal differentiation and survival and synaptic plasticity suggest that NO/cGMP regulate gene expression in the nervous system (1– 6). Regulation of gene expression by the NO/cGMP signal transduction pathway has been recognized only recently, with induction of c-fos mRNA by cGMP-elevating agents and membranepermeable cGMP analogs occurring in a variety of cell types, including fibroblasts, epithelial, glial, and neuronal cells (20 –25). In some cell types, NO and cGMP appear to induce c-fos mRNA only when cells 1 Correspondence: Department of Medicine, University of California, San Diego, 9500 Gilman Dr., La Jolla, CA 920930652. E-mail: [email protected]

2143

are costimulated with agents that increase the intracellular calcium concentration (23, 25). We have shown in fibroblast-like cells that G-kinase I mediates cGMP activation of the fos promoter and that transcriptional regulation requires nuclear translocation of the kinase (26, 27). We identified a nuclear localization signal in G-kinase I that is both necessary for nuclear translocation of the enzyme and sufficient to direct transport of a heterologous protein to the nucleus (27). Nuclear translocation of G-kinase I occurs only after activation by cGMP, suggesting that the nuclear localization signal is cryptic in the inactive enzyme and is exposed through a cGMP-induced conformational change or that cGMP changes interactions of the kinase with other proteins (27, 28). G-kinase II contains a sequence [amino acids 482– 489 of rat G-kinase II (29)] that closely resembles the nuclear localization signal identified in G-kinase I (27), but it is not known whether membrane-associated G-kinase II can change its subcellular localization and/or whether G-kinase II regulates gene expression in response to cGMP activation. Our findings establish G-kinase II as a mediator of transcriptional regulation in cells of neuronal and glial origin and have important physiological implications with regard to NO, cGMP, and calcium signaling in the brain.

Subcellular fractionation and Western blot analyses Cells were lysed in 0.4 ml of buffer A (150 mM NaCl, 10 mM NaPO4, 1 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride, 20 mg/ml leupeptin, 10 mg/ml aprotinin, and 100 mg/ml soybean trypsin inhibitor, pH 7.4) using a Dounce homogenizer (20 strokes, type B pestle) and brief sonication (three bursts of 3 s). Cytosol and particulate fractions were separated by centrifugation at 100,000 3 g for 1 h at 4°C. Whole cell homogenates and cytosolic and particulate fractions corresponding to 104 cells each were analyzed by denaturing polyacrylamide gel electrophoresis/Western blotting using previously characterized polyclonal antibodies against G-kinase I (1:2000 dilution) or G-kinase II (1:1000 dilution) and enhanced chemiluminescence detection (30, 33). Forebrains of 3-day-old pups of Sprague-Dawley rats were dissected and snap-frozen in the laboratory of Dr. J. A. Frangos of the University of California at San Diego. The tissue was homogenized and sonicated in buffer A, as described above, and half of the homogenate was incubated for 5 min at 4°C with 0.5 M NaCl and 1% Triton X-100 (16). After centrifugation at 100,000 3 g for 1 h, the supernatants were preabsorbed with protein G agarose and incubated with G-kinase II-specific antibody (1:200 dilution) and protein G agarose for 3 h at 4°C. Washed beads were analyzed by Western blots developed using G-kinase II antibody. Protein kinase assays Cells were transfected as described above and solubilized for 5 min at 4°C with 1% Triton X-100 in buffer A in the presence of 0.5 M NaCl (16). Lysates were cleared by centrifugation for 5 min at 11,000 3 g and total cellular G-kinase activity was determined in the presence of the specific protein kinase A inhibitor PKI, as described (26).

MATERIALS AND METHODS Immunofluorescence staining of G-kinase Plasmids The expression vectors for wild-type G-kinase Ib and G-kinase II, the G-kinase II (G2- A) mutant, the G-kinase II/I chimera that contains the first 29 amino acids of G-kinase II fused to the NH2 terminus of G-kinase I, and the G-kinase I (K407-E) mutant have been described (16, 27, 29 –31). The chloramphenicol acetyltransferase (CAT) reporter gene under control of the human fos promoter (pFos-CAT) and the reporter pTF5/53 were also described previously (26, 32).

Cells were transfected and fixed as described (27). After blocking with 5% bovine serum albumin, cells were incubated with antibody against G-kinase I or II (dilution 1:500), followed by fluorescein-5-isothiocyanate-conjugated goat antirabbit immunoglobulin G (Cappel), and visualized with a Zeiss fluorescence microscope (27).

RESULTS

Transient transfections and reporter gene assays

Cell type-specific trans-activation of a cGMPresponsive reporter by G-kinase II

Baby hamster kidney (BHK) cells, NIH3T3, C6 glioma, and PC12 pheochromocytoma cells were transfected using Lipofectamine (Life Technologies, Grand Island, N.Y.) as described previously (26); NB2A murine neuroblastoma cells were transfected similarly using Fugene (Boehringer Mannheim Corp., Mannheim, Germany). After transfection, cells were cultured in low serum-containing medium (26) and treated as indicated with 8-para-chlorophenylthio-cGMP (8CPT-cGMP, BioLog), ethanamine,2,29-(hydroxynitroso-hydrazono)bis- (Deta NONOate, Cayman), or A23187 (Sigma, St. Louis, Mo.) for 8 h prior to harvesting; 1H-[1, 2, 4]oxadiazolo[4.3-a]quinoxalin-1-one (ODQ, Calbiochem) was added 24 h before harvesting, where indicated. CAT and luciferase activities were measured as described (26). Transfection efficiencies were determined by transfection of a b-galactosidase expression vector and histochemical staining (27).

Although G-kinase II is found in a variety of tissues, it has not been reported to be expressed in cultured cells, presumably because enzyme expression is lost during in vitro culture. We screened several primary cell cultures and more than 30 established cell lines including neuronal, intestinal epithelial, and kidneyderived cell lines and found no detectable G-kinase II expression by Western blotting. (T. Gudi, S. M. Lohmann and R. B. Pilz, unpublished results). Although we found small amounts of G-kinase I in early passage C6 rat glioma and NB2A murine neuroblastoma cells, the amount of G-kinase I declined to negligible levels after several passages, as de-

2144

Vol. 13

December 1999

The FASEB Journal

GUDI ET AL.

TABLE 1. G-kinase activity in C6 glioma and BHK cells transfected with G-kinase constructs G-kinase activityb G-kinase construct transfecteda

C6 glioma cellsc

BHK cellsc

[nmol/(min z mg protein)] Mock transfected G-kinase I (wild-type) G-kinase II/I chimera G-kinase II (wild-type) G-kinase II (G2-A)

0.02 6 0.02 0.72 6 0.11 1.32 6 0.16 0.72 6 0.16 0.69 6 0.10

0.01 6 0.01 1.08 6 0.13 1.95 6 0.11 1.15 6 0.16 0.49 6 0.02

a Cells were transfected with the indicated G-kinase constructs as b described in Materials and Methods. G-kinase activity was measured as described in Materials and Methods; values are the mean 6 sd of three independent experiments performed in duplicate (values c are not corrected for transfection efficiencies). Transfection efficiency was determined as described in Materials and Methods and was 45–50% for C6 glioma cells and 75– 80% for BHK cells.

scribed for other cell types (34). To compare the function of G-kinase II with that of G-kinase I in the same cellular environment, we transfected the enzymes into different cell types that are almost completely G-kinase deficient by activity assay and Western blot analysis (Table 1; Fig. 4A;27). We previously found that the most cGMP-responsive cis-acting element in the human fos promoter is between nucleotides -303 and –281, which includes the fos AP-1 binding site (26, 32). We used a reporter containing this enhancer element upstream of the fos basal promoter to examine the capacity of Gkinase II to regulate gene expression in different cell types; cotransfection of a luciferase expression vector served as an internal control for transfection efficiency. In C6 rat glioma cells, PC12 rat pheochromocytoma cells, and NB2A murine neuroblastoma cells transfected with G-kinase II, the membrane-permeable cGMP analog 8-CPT-cGMP increased reporter gene expression four- to sixfold (Fig. 1). This is in contrast to BHK cells and NIH 3T3 fibroblasts, where G-kinase II had only a small effect or no effect on reporter gene expression, respectively (Fig. 1). When cells were transfected with G-kinase Ib, 8-CPTcGMP increased reporter gene expression four- to eightfold in all cell types examined (Fig. 1). In cells transfected with empty vector, 8-CPT-cGMP had no significant effect on reporter gene expression, indicating little endogenous G-kinase activity and lack of cross-activation of cAMP-dependent protein kinase (A-kinase) by 8-CPT-cGMP (Fig. 1). For further studies of G-kinase II regulation of gene expression, we chose C6 glioma cells and BHK cells that represent cell types with high and low G-kinase II-mediated activation of gene expression, respectively. Similar to the findings described above, the full-length fos promoter was stimulated ;fourfold by 8-CPT-cGMP in G-kinase II-expressing C6 cells, whereas 8-CPT-cGMP had no significant effect

in BHK cells expressing this kinase (see Fig. 6A, B). For comparison, the full-length fos promoter was stimulated four- to fivefold by 8-CPT-cGMP in both C6 cells and BHK cells expressing G-kinase I (see Fig. 6A, B). NO induction of the fos promoter requires G-kinase NO, by itself, has been reported to induce c-fos mRNA expression in different cell types, including PC12 pheochromocytoma cells (21, 22, 35–37); however, other workers have found in PC12 cells and GT1 hypothalamic cells that NO required increased intracellular calcium to induce c-fos mRNA (23, 25). To determine whether NO induction of c-fos was mediated by G-kinase, we treated G-kinase-deficient and G-kinase II-expressing C6 cells with the NOgenerating agent Deta NONOate. In G-kinase-deficient C6 cells, transcription from the full-length fos promoter was not significantly increased by NO, but in cells expressing G-kinase II, fos promoter activity was stimulated three- to fourfold by 250 mM Deta NONOate (Fig. 2). Similar results were obtained with G-kinase I-expressing C6 cells (data not shown). Preincubation of cells with the calcium chelator EGTA (5 mM) did not influence NO- or cGMP-

Figure 1. Transactivation of a cGMP-responsive reporter by G-kinase I and II in different cell types. Cells were transfected with 50 ng of pTF/53 (a cGMP-responsive reporter), 50 ng of pRSV-Luc (internal control), and either 500 ng of the indicated G-kinase expression vector or control vector (pRC/ CMV). Cells were either left untreated or were treated with 500 mM 8-CPT-cGMP (cGMP), as indicated, for 8 h prior to measuring reporter gene activities. CAT activity was normalized to luciferase activity in each sample and expressed relative to the CAT/luciferase ratio of untreated cells transfected with control vector, which was assigned a value of 1 (open bars). Cell lines: C6, rat glioma cells; PC12, rat pheochromocytoma cells; NB2A, murine neuroblastoma cells; BHK, baby hamster kidney cells; NIH3T3, murine fibroblasts. Data are the mean 6 sd of at least three independent experiments.

TRANSCRIPTIONAL REGULATION BY cGMP-DEPENDENT PROTEIN KINASE II

2145

Figure 2. Effects of NO and calcium on the fos promoter in G-kinase-deficient and G-kinase-expressing C6 glioma cells. Cells were transfected with pFosCAT, pRSV-Luc, and an expression vector encoding wild-type G-kinase II or empty control vector; reporter gene activities were measured as described. Cells were treated with 250 mM 8-CPT-cGMP, 250 mM Deta NONOate, or 0.3 mM A23187 for 8 h prior to harvesting; in some cases, cells were pretreated with the soluble guanylate cyclase inhibitor ODQ (10 mM) for 16 h as indicated. Inset: Cells transfected with empty vector (squares) or G-kinase II expression vector (triangles) were treated for 8 h with increasing concentrations of A23187. Some of the G-kinase-transfected cells were treated only with the indicated concentrations of A23187 (open triangles), others were also treated with 250 mM 8-CPT-cGMP (closed triangles). Data are the mean 6 sd of three independent experiments performed in duplicate.

induced fos promoter activity in G-kinase-expressing cells, whereas this concentration of EGTA completely prevented calcium ionophore effects on the fos promoter (data not shown). These results indicate that stimulation of the fos promoter by NO required the presence of G-kinase but that the effects of NO did not depend on increased extracellular calcium uptake. Synergistic activation of the fos promoter by NO/ cGMP and calcium requires G-kinase We used the calcium ionophore A23187 to test whether the effect of NO and cGMP on the fos promoter could be augmented by increasing the intracellular calcium concentration. Treating G-kinase-deficient C6 cells with A23187 increased fos promoter activity modestly, as described in other cell types [23; Fig. 2: the effect of 0.3 mM A23187 is shown in the bar graph and the effects of varying A23187 concentrations are shown in the inset, (open 2146

Vol. 13

December 1999

squares)]. Calcium stimulation of the fos promoter is mediated by calcium/calmodulin-dependent protein kinases (Cam-kinases) and the cAMP-responsive element binding protein CREB (38, 39). Adding Deta NONOate or 8-CPT-cGMP with A23187 had no significant effect in the G-kinase-deficient C6 cells (Fig. 2, bar graph). In G-kinase II-expressing C6 cells, A23187 increased fos promoter activity significantly more than in G-kinase-deficient cells, suggesting that G-kinase II was activated under these conditions (Fig. 2, inset: compare open squares and open triangles, representing G-kinase-deficient and G-kinase II-expressing cells, respectively). When G-kinase II-expressing C6 cells were treated with the soluble guanylate cyclase inhibitor ODQ prior to adding A23187, the ionophore’s effect on the fos promoter was reduced to that observed in G-kinase-deficient cells (Fig. 2, bar graph); 8-CPT-cGMP reversed the effect of ODQ (data not shown). These results suggest that increasing the intracellular calcium concentration in C6 cells induced NO synthesis, leading to increased intracellular cGMP and G-kinase activation (2, 40). Thus, in G-kinase-expressing C6 cells, the effect of A23187 on the fos promoter was mediated in part by activation of NO synthase, guanylate cyclase, and G-kinase in addition to activation of Cam-kinases. It appeared that G-kinase II was only partially activated by the endogenous NO/cGMP produced in A23187treated cells, because the addition of Deta NONOate or 8-CPT-cGMP to these cells further increased fos promoter activity (Fig. 2, bar graph and inset). Since calcium induced the fos promoter only modestly in G-kinase-deficient C6 cells and in G-kinaseexpressing cells treated with the guanylate cyclase inhibitor ODQ, it appears that NO/cGMP and calcium acted synergistically to stimulate the fos promoter, as described previously in PC12 cells (23). Our results demonstrate that this synergistic action of NO/cGMP and calcium required G-kinase. Similar, albeit less pronounced, synergy between NO/ cGMP and calcium was observed in G-kinase I-expressing C6 cells (data not shown). Comparison of G-kinase I and II expression and subcellular localization The difference in the ability of G-kinase I and II to regulate gene expression in C6 glioma cells and BHK cells (see Figs. 1 and 6) could be due to differences in expression levels or subcellular localization of the enzymes. When transfection efficiencies were accounted for, G-kinase activity was similar in C6 cells and BHK cells transfected with each respective G-kinase expression vector and comparable to specific activities reported in tissues that express the endogenous enzymes (Table 1; 9,

The FASEB Journal

GUDI ET AL.

15, 26). When the subcellular distribution of Gkinase I and II was examined in transfected C6 and BHK cells, 50 – 80% of G-kinase I was in the 100,000 3 g supernatant, whereas G-kinase II was found almost exclusively (.90%) in the pellet (Figs. 4A and 5B show results in C6 and BHK cells, respectively). This distribution of G-kinase I and II between soluble and particulate fractions resembles that of endogenous G-kinase I in smooth muscle and endothelial cells and G-kinase II in kidney and intestinal epithelial cells (9, 33, 41; T. Gudi, unpublished results). G-kinase II was also particulate in rat brain: it was detectable in immunoprecipitates of 100,000 3 g supernatants only after solubilization with salt and detergent (Fig. 3). Mouse brain G-kinase II expressed in COS cells was reported to reside partially in the cytosol, but this finding might be due to high expression levels saturating membrane binding sites or to an aminoterminal poly-histidine tag interfering with membrane association of the enzyme (14, 42). We have previously demonstrated that G-kinase I translocates to the nucleus upon activation by cGMP (27). Membrane association of G-kinase II does not preclude its nuclear translocation because membrane association of some proteins is reversible (43). We therefore examined whether G-kinase II changes its subcellular localization when cells are treated with 8-CPT-cGMP. Indirect immunofluorescence staining of G-kinase II in C6 glioma and BHK cells showed nonhomogeneous distribution of the enzyme, with intense staining of the cell periphery and the perinuclear region (Fig. 4B and Fig. 5C show results in C6 and BHK cells, respectively). In G-kinase II-express-

Figure 3. Immunoprecipitation of G-kinase II from detergentsolubilized rat brain. Forebrains of 3-day-old rats were homogenized and half of the homogenate was treated with 0.5 M NaCl and 1% Triton-X-100, as described in Materials and Methods. After centrifugation at 100,000 3 g, supernatants were subjected to immunoprecipitation using a G-kinase II-specific antibody. Immunoprecipitates corresponding to one-tenth of one forebrain (lanes 1 and 3) were applied to denaturing polyacrylamide electrophoresis side-by-side with extracts of G-kinase II-expressing C6 glioma cells (lanes 2 and 4); Western blots were made using a G-kinase II antibody, as described in Materials and Methods. Shown are the immunoprecipitates of 100,000 3 g supernatants of a salt- and detergent-treated homogenate (lane 1) and of a control homogenate (lane 3); G-kinase II-expressing C6 glioma cell extracts served as standards (lanes 2 and 4).

Figure 4. Expression and subcellular localization of G-kinase I and II in C6 glioma cells. A) Western blot analysis. Cells were transfected with expression vectors encoding either wild-type G-kinase I [GK I (WT)] or G-kinase II [GK II (WT)] as described in Materials and Methods. Cell homogenates (H) and 100,000 3 g supernatant (S) or pellet (P) fractions, corresponding to 104 cells each, were analyzed by Western blotting using an antibody specific for G-kinase I (left panel) or G-kinase II (right panel). Mock-transfected C6 cells were processed in parallel. B) Immunofluorescence analysis. Cells were transfected with G-kinase I and II expression vectors; 36 h after transfection, some of the cultures were treated with 500 mM 8-CPT-cGMP for 1 h as indicated (right panels) prior to staining with an antibody specific for G-kinase I (upper panels) or G-kinase II (lower panels). Indirect immunofluorescence staining was performed as described in Materials and Methods. No fluorescence signal was observed in mocktransfected cells (not shown).

ing cells treated with 8-CPT-cGMP, no significant nuclear staining was observed; this is in contrast to cells expressing G-kinase I, which show nuclear translocation of the kinase under these conditions (Fig. 4B and Fig. 5C). These differences between G-kinase I and II are unlikely to be secondary to differential enzyme activation by 8-CPT-cGMP because the affinity of G-kinase II for 8-CPT-cGMP is higher than that of G-kinase Ib (15, 44). We cannot exclude that a small amount of soluble G-kinase II (undetectable by immunofluorescence) translocated to the nucleus; however, this could not account for the similar degree of transcriptional activation observed for G-kinase I and II in C6 glioma cells (Fig. 1 and Fig. 6A).


2147

Figure 5. Expression and subcellular localization of G-kinase I and II variants in BHK cells. A) Schematic representation of G-kinase I and II constructs: GKI (WT), wild-type G-kinase Ib; GK II (WT), wild-type G-kinase II with amino-terminal myristoylation; GKII (G2-A), G-kinase II with a glycine to alanine substitution at position two that prevents myristoylation; GKII/I chimera, a chimera that contains the first 29 amino acids of G-kinase II fused to the NH2 terminus of G-kinase I. B) Western blot analysis. Cells were transfected with the indicated G-kinase expression vector; cell homogenates (H) and 100,000 3 g supernatant (S) or pellet (P) fractions were analyzed as described in Fig. 4A using an antibody specific for G-kinase I (left panel) or G-kinase II (right panel). No detectable G-kinase I or II expression was observed in equivalent samples from mocktransfected BHK cells (ref 27 and data not shown). C) Immunofluorescence analysis. Cells were transfected with the G-kinase expression vectors described above. At 36 h after transfection, half of the cultures were treated with 500 mM 8-CPT-cGMP for 1 h as indicated (right panels) prior to staining with an antibody specific for G-kinase I (upper four panels) or G-kinase II (lower four panels). Indirect immunofluorescence staining was performed as described in Materials and Methods. No fluorescence signal was observed in mock-transfected cells (not shown).

Expression and subcellular localization of G-kinase I and II variants To determine whether membrane association of G-kinase II prevents nuclear localization of the enzyme, we examined the subcellular localization of two variant G-kinase constructs: a mutant G-kinase II with a glycine to alanine substitution at position two that prevents myristoylation [GK II (G2-A) mutant] and a membrane-targeted variant of G-kinase I that contains the first 29 amino acids of G-kinase II fused to the NH2 terminus of G-kinase I [G-kinase II/I chimera ] (Fig. 5A). Approximately 50% of the G-kinase II (G2-A) mutant was found in the soluble fraction of C6 and BHK cells (Fig. 5B, right panel, shows wild-type and mutant G-kinase II in BHK cells). These results suggest that G-kinase II may be targeted to membranes or subcellular organelles in C6 and BHK cells by additional mechanisms besides myristoylation, e.g., by interaction with anchoring 2148

Vol. 13

December 1999

proteins (45). More than 90% of the G-kinase II/I chimera was in the particulate fraction of C6 and BHK cells (Fig. 5B, left panel, shows wild-type Gkinase I and the G-kinase II/I chimera in BHK cells). These results agree with previous work demonstrating that the first 29 amino acids of G-kinase II are sufficient to direct G-kinase I to the plasma membrane of intestinal epithelial cells (31). The higher expression of the G-kinase II/I chimera compared to wild-type G-kinase I (Fig. 5B and Table 1) may be because the amino-terminal myristoylation signal stabilizes the enzyme (16). Indirect immunofluorescence staining demonstrated that in the absence of 8-CPT-cGMP, the partially soluble G-kinase II (G2-A) mutant was diffusely distributed throughout cells with sparing of the nuclei; when cells were treated with 8-CPTcGMP, there was prominent nuclear staining (Fig. 5C), which was confirmed by confocal laser microscopy (not shown). In 8-CPT-cGMP-treated cells expressing the membrane-targeted G-kinase II/I chimera, no nuclear staining was demonstrated (Fig. 5C). The data suggest that G-kinase II contains a functional nuclear localization signal that is exposed

The FASEB Journal

GUDI ET AL.

when the kinase is activated by cGMP, but that membrane association of G-kinases I or II via the first 29 amino acids, including the myristoylation signal of G-kinase II, is sufficient to prevent nuclear translocation of the kinases. Transactivation of the fos promoter by G-kinase I and II variants Since the partially soluble G-kinase II (G2-A) mutant resembled G-kinase I and the membrane-targeted G-kinase II/I chimera resembled wild-type G-kinase II with respect to subcellular localization and capacity for cGMP-induced nuclear translocation (Fig. 5), we examined the effect of these G-kinase variants on fos promoter activity in C6 glioma and BHK cells. The partially soluble G-kinase II (G2-A) mutant mediated an ;sixfold induction of the fos promoter in both cell types treated with 8-CPT-cGMP (Fig. 6A, B). This induction correlated with G-kinase II (G2-A) translocation to the nucleus (Fig. 5C) and suggested that nuclear translocation of G-kinase is required for efficient transcriptional regulation in BHK cells. Since G-kinase II (G2-A) induced the fos promoter more efficiently than either wild-type G-kinase I or II in C6 cells, the mutant enzyme may recognize nuclear substrates more efficiently than G-kinase I and/or may act through both nuclear and extranuclear mechanisms. As expected from its lack of nuclear translocation in immunofluorescence studies (Fig. 5C), the membrane-targeted G-kinase II/I chimera, like wild-type G-kinase II, was unable to activate the fos promoter in BHK cells (Fig. 6B). The G-kinase II/I chimera also failed to trans-activate the fos promoter in C6 glioma cells, suggesting that the factor(s) that mediate the nuclear effects of membrane-bound G-kinase II in these cells are not recognized by this membranetargeted G-kinase I variant (Fig. 6A). Membrane targeting could potentially alter substrate recognition by the G-kinase II/I chimera, but the enzyme is fully active in vitro toward several substrates and phosphorylates the cystic fibrosis transmembrane conductance regulator (CFTR) channel protein in vivo (31). To test another extranuclear G-kinase I construct, we examined G-kinase I (K407-E), which is unable to translocate to the nucleus in response to cGMP due to a mutation in the nuclear localization signal (27). This mutant G-kinase I (K407-E) is indistinguishable from wild-type G-kinase I with respect to solubility, in vitro kinase activity, cGMP binding, and in vivo phosphorylation of the extranuclear substrate vasodilator-stimulated phosphoprotein (27; T. Gudi, unpublished results). In C6 and BHK cells transfected with G-kinase I (K407-E), the fos promoter was unresponsive to cGMP (Fig. 6A, B). The finding that two different versions of G-kinase I

Figure 6. Transactivation of the fos promoter by G-kinase I and II variants. Cells were transfected with 50 ng of pFosCAT, 50 ng of pRSV-Luc (internal control), and 500 ng of the indicated G-kinase expression vector or empty control vector, as described in Materials and Methods. Cells were either left untreated (striped bars) or were treated with 500 mM 8-CPTcGMP (filled bars) for 8 h prior to measuring reporter gene activities. CAT activity was normalized to luciferase activity as described in Fig. 1. The G-kinase constructs are described in Fig. 5 except for GK I (K407-E), which encodes G-kinase I with a mutation in the nuclear localization signal that prevents nuclear translocation (27). A) C6 glioma cells; B) BHK cells. Data are the mean 6 sd of three independent experiments.

that are excluded from the nucleus [G-kinase II/I chimera and G-kinase (K407-E)] did not activate the fos promoter in C6 glioma cells strongly suggests that extranuclear factor(s) that mediate trans-activation by G-kinase II are not recognized by G-kinase I. DISCUSSION Many of the functions of NO and cGMP in the nervous system may involve regulation of gene expression; e.g., up-regulation of mRNAs encoding c-fos and microtubule-associated protein-2 and down-regulation of the mRNAs encoding gonadotropin-re-


2149

leasing hormone and calcium/calmodulin-dependent protein kinase IIa (1, 20, 23, 25, 46, 47). We have previously demonstrated that trans-activation of the human fos promoter by NO and cGMP is mediated by G-kinase I in BHK cells (26, 48). However, the distribution of G-kinase I in the nervous system is quite limited, with high levels found only in cerebellar Purkinje cells and sensory neurons of the dorsal root ganglia (1, 13, 18). In contrast, G-kinase II is expressed in most regions of the brain, with highest levels in the thalamus, cortex, septum, amygdala, and olfactory bulb, and low levels in cerebellum, striatum, and hippocampus (12, 17). We found that G-kinase II trans-activated the human fos promoter as efficiently as G-kinase I in cells of neuronal and glial origin, but that transcriptional regulation by G-kinase II was cell type specific. Unlike G-kinase I, G-kinase II did not require nuclear translocation of the kinase to trans-activate the fos promoter. Moreover, the factor(s) mediating the nuclear effects of G-kinase II in C6 cells did not appear to be recognized by G-kinase I, since two different extranuclear versions of G-kinase I failed to trans-activate the fos promoter in these cells. Differences between G-kinase I and II substrate recognition have been reported (14, 15); in the case of the CFTR channel, this difference could be overcome by targeting G-kinase I to the same subcellular compartment as G-kinase II (31). Differential access of G-kinase I and II to specific substrates in vivo may be determined not only by membrane targeting of the latter enzyme, but also by the presence or absence of cell type-specific G-kinase II anchoring proteins, which have been described recently (45). Protein kinase A is targeted to different subcellular compartments via its association with specific anchoring proteins, and some physiological functions of protein kinase A require association of the kinase with specific anchoring proteins (49). The cell-specific factors that determine the differential capacity of G-kinase II to regulate gene expression in different cell types may include other protein kinases or phosphatases regulated by G-kinase II and/or transcription factors that translocate to the nucleus in response to phosphorylation. Preliminary work from our laboratory indicates that activation of the fos promoter by G-kinase II in C6 glioma cells does not involve activation of the mitogen-activated protein kinase pathway. Several specific peptide inhibitors of protein phosphatase-1 expressed in neuronal tissues are activated by G-kinase phosphorylation (50); inhibition of protein phosphatase-1 can lead to activation of the fos promoter (51, 52). A candidate transcription factor that could be regulated by G-kinase II phosphorylation is the CCAAT/ enhancer binding protein (C/EBP)-b; this protein is expressed in cells of neuronal and glial origin and 2150

Vol. 13

December 1999

translocates to the nucleus in response to phosphorylation by protein kinase A (53, 54). Work is in progress to examine the effect of G-kinase I and II on subcellular localization and function of C/EBP-b. Preliminary experiments demonstrate that trans-activation of the fos promoter by G-kinase I and II requires similar cis-acting elements (T. Gudi et al., unpublished results). Our findings are of particular relevance to the function of the NO/cGMP signal transduction pathway in the regulation of synaptic plasticity (1– 4, 7, 8). Long-term changes in synaptic efficiency, i.e., longterm potentiation and depression (LTP/LTD), are thought to underlie the processes of learning and memory and require activation of new programs of gene expression (reviewed in refs 38, 39). Hippocampal LTP is reduced in mice that are doubly mutant in endothelial and neuronal NO synthase, indicating that NO is critical for facilitated synaptic transmission (55); however, hippocampal LTP appears to be normal in mice lacking type I and II G-kinases, suggesting that NO induces LTP through an alternative pathway (56). In contrast, G-kinase seems to be involved in the regulation of LTD (2, 3, 8). Simultaneous increases in the intracellular calcium and NO/cGMP concentrations are both necessary and sufficient for induction of cerebellar LTD; thus, synergistic regulation of gene expression by calcium and NO/cGMP in neuronal cells may be important for the regulation of LTD (2, 3, 23). Our results demonstrate that the effect of NO and the synergistic effect of NO and calcium on the fos promoter in C6 cells require G-kinase activity; the synergism between NO and calcium may involve G-kinase phosphorylation of protein phosphatase-1 inhibitors, because inhibition of protein phosphatase-1 can lead to increased Cam-kinase activation and CREB phosphorylation (50, 52, 57). Until now, only indirect evidence existed for Gkinase regulating gene expression in the nervous system. NO-releasing agents and membrane-permeable cGMP analogs increase c-fos mRNA expression in calcium ionophore-treated PC12 pheochromocytoma and GT1 hypothalamic cells (23, 25): in the former study, c-fos induction was reduced by an protein kinase A inhibitory peptide that likely also inhibited G-kinase activity; in the latter study, the effect was prevented by a specific G-kinase inhibitor. Other workers have reported induction of c-fos mRNA by NO and cGMP analogs in PC12 cells in the absence of calcium costimulation, and inhibition of the effect by a G-kinase inhibitor (35). PC12 cell lines may vary in G-kinase activity and thus in NOinduced increases in c-fos mRNA; the PC12 cell line we have examined is G-kinase deficient and does not show c-fos mRNA induction in response to NO or 8-CPT-cGMP (T. Gudi and R. B. Pilz, unpublished

The FASEB Journal

GUDI ET AL.

results). NO and cGMP can also modulate the expression of c-fos mRNA and other genes in various primary neuronal tissues; in some cases, G-kinase inhibitors were shown to abrogate the effect (36, 37, 46, 47, 58, 59). Unfortunately, G-kinase inhibitors do not allow distinction between G-kinase I and II (13, 20). Although different amounts of G-kinase I and II are found in specific neuronal cell subpopulations, both enzymes are coexpressed in some cell types, e.g., in cerebellar and hippocampal neurons (1, 11, 12, 17, 18). We took advantage of transient transfection assays in G-kinase-deficient cell lines to examine the effect of G-kinase I and II independently. Our findings clearly indicate that regulation of gene expression by the membrane-bound, extranuclear G-kinase II requires cell-specific factors that are found in some cells of neuronal and glial origin. In contrast, G-kinase I regulates gene expression in a wide variety of cell types, but transcriptional regulation requires nuclear translocation of the kinase.

10.

11.

12.

13.

14.

15.

16.

We thank S. R. P. Gudi and J. A. Frangos for providing rat brain tissue, C. Buckmaster and J. Feramisco for assistance with fluorescence microscopy, S. R. P. Gudi and G. R. Boss for helpful discussions, and V. S. Sharma for his support and active intellectual participation. T.G. was supported in part by American Heart Association Grant-in-Aid #9650582N and NIH grant 2T32DK07233, S.M.L. was supported by the Deutsche Forschungsgemeinschaft grant SFB 355, A.B.V. was supported by the Netherlands Organization for Scientific Research (NWO) and R.B.P. was supported by National Science Foundation grant MCB-9506327 and NIH grant 1RO1-GM055586.

REFERENCES 1. 2. 3.

4.

5.

6. 7. 8.

9.

Wang, X, and Robinson, P. J. (1997) Cyclic GMP-dependent protein kinase and cellular signaling in the nervous system. J. Neurochem. 68, 443– 456 Garthwaite, J., and Boulton, C. L. (1995) Nitric oxide signaling in the central nervous system. Annu. Rev. Physiol. 57, 683–706 Lev-Ram, V., Jiang, T., Wood, J., Lawrence, D. S., and Tsien, R. Y. (1997) Synergies and coincidence requirements between NO, cGMP, and Ca21 in the induction of cerebellar long-term depression. Neuron 18, 1025–1038 Zhuo, M., Yinghe, H., Schultz, C., Kandel, E. R., and Hawkins, R. D. (1994) Role of guanylyl cyclase and cGMP-dependent protein kinase in long-term potentiation. Nature (London) 368, 635– 639 Farinelli, S. E., Park, D. S., and Greene, L. A. (1996) Nitric oxide delays the death of trophic factor-deprived PC12 cells and sympathetic neurons by a cGMP-mediated mechanism. J. Neurosci. 16, 2325–2334 Peunova, N., and Enikolopov, G. (1995) Nitric oxide triggers a switch to growth arrest during differentiation of neuronal cells. Nature (London) 375, 68 –73 Lewin, M. R., and Walters, E. T. (1999) Cyclic GMP pathway is critical for inducing long-term sensitization of nociceptive sensory neurons. Nature (London) Neuroscience 2, 18 –23 Calabresi, P., Gubellini, P., Centonze, D., Sancesario, G., Morello, M., Giorgi, M., Pisani, A., and Bernardi, G. (1999) A critical role of the nitric oxide/cGMP pathway in corticostriatal long-term depression. J. Neurosci. 19, 2489 –2499 Butt, E., Geiger, J., Jarchau, T., Lohmann, S. M., and Walter, U. (1993) The cGMP-dependent protein kinase - gene, protein, and function. Neurochem. Res. 18, 27– 42

17.

18.

19.

20.

21.

22.

23. 24.

25.

26. 27.

28.

Lincoln, T. M. (1995) Cyclic GMP receptor proteins: role in nervous system and other tissues. In Nitric Oxide in the Nervous System (Vincent, S. R., ed ) pp. 51–91, Academic Press, San Diego, California Li, H., Liu, J. -P., Smith, R., and Robinson, P. J. (1996) Identification of cGMP-dependent protein kinase and its specific substrates in the anterior pituitary. Mol. Cell Endocrinol. 122, 159 –171 El-Husseini, A. E. D., Bladen, C., Williams, J. A., Reiner, P. B., and Vincent, S. R. (1998) Nitric oxide regulates cyclic GMPdependent protein kinase phosphorylation in rat brain. J. Neurochem. 71, 676 – 683 Lohmann, S. M., Vaandrager, A. B., Smolenski, A., Walter, U., and De Jonge, H. R. (1997) Distinct and specific functions of cGMP-dependent protein kinases. Trends Biochem. Sci. 22, 307– 312 Gamm, D. M., Francis, S. H., Angelotti, T. P., Corbin, J. D., and Uhler, M. D. (1995) The type II isoform of cGMP-dependent protein kinase is dimeric and possesses regulatory and catalytic properties distinct from the type I isoforms. J. Biol. Chem. 270, 27380 –27388 Vaandrager, A. B., Edixhoven, M., Bot, A. G. M., Kroos, M. A., Jarchau, T., Lohmann, S., Genieser, H. -G., and De Jonge, H. R. (1997) Endogenous type II cGMP-dependent protein kinase exists as a dimer in membranes and can be functionally distinguished from the type I isoforms. J. Biol. Chem. 272, 11816 – 11823 Vaandrager, A. B., Ehlert, E. M. E., Jarchau, T., Lohmann, S. M., and De Jonge, H. R. (1996) N-terminal myristoylation is required for membrane localization of cGMP-dependent protein kinase type II. J. Biol. Chem. 271, 7025–7029 El-Husseini, A. E. D., Bladen, C., and Vincent, S. R. (1995) Molecular characterization of a type II cyclic GMP-dependent protein kinase expressed in the rat brain. J. Neurochem. 64, 2814 –2817 Qian, Y., Chao, D. S., Santillano, D., Cornwell, T. L., Nairn, A. C., Greengard, P., Lincoln, T. M., and Bredt, D. S. (1996) cGMP-dependent protein kinase in dorsal root ganglion: relationship with nitric oxide synthase and nociceptive neurons. J. Neurosci. 16, 3130 –3138 Orstavik, J., Natarajan, V., Tasken, K., Jahnsen, T., and Sandberg, M. (1997) Characterization of the human gene encoding the type Ia and type Ib cGMP-dependent protein kinase (PRKG1). Genomics 42, 311–318 Eigenthaler, M., Lohmann, S. M., Walter, U., and Pilz, R. B. (1998) Signal transduction by cGMP-dependent protein kinases and their emerging roles in the regulation of cell adhesion and gene expression. Rev. Physiol. Biochem. Pharmacol. 135, 174 –209 Pilz, R. B., Suhasini, M., Idriss, S., Meinkoth, J. L., and Boss, G. R. (1995) Nitric oxide and cGMP analogs activate transcription from AP-1-responsive promoters in mammalian cells. FASEB J. 9, 552–558 Bosco-Felley, E., Ambs, S., Lowenstein, C. J., Keefer, L. K., and Harris, C. C. (1994) Constitutive expression of inducible nitric oxide synthase in human bronchial epithelial cells induces c-fos and stimulates the cGMP pathway. Am. J. Resp. Cell Mol. Biol. 11, 159 –164 Peunova, N., and Enikolopov, G. (1993) Amplification of calcium-induced gene transcription by nitric oxide in neuronal cells. Nature (London) 364, 450 – 453 Thiriet, N., Esteve, L., Aunis, D., and Zwiller, J. (1997) Immediate early gene induction by natriuretic peptides in PC12 phaeochromocytoma and C6 glioma cells. Mol. Neurosci. 8, 399 – 402 Belsham, D. D., Wetsel, W. C., and Mellon, P. L. (1996) NMDA and nitric oxide act through the cGMP signal transduction pathway to repress hypothalamic gonadotropin-releasing hormone gene expression. EMBO J. 15, 538 –547 Gudi, T., Huvar, I., Meinecke, M., Lohmann, S. M., Boss, G. R., and Pilz, R. B. (1996) Regulation of gene expression by cGMPdependent protein kinase. J. Biol. Chem. 271, 4597– 4600 Gudi, T., Lohmann, S. M., and Pilz, R. B. (1997) Regulation of gene expression by cGMP-dependent protein kinase requires nuclear translocation of the kinase: identification of a nuclear localization signal. Mol. Cell. Biol. 17, 5244 –5254 Chu, D. -M., Corbin, J. D., Grimes, K. A., and Francis, S. H. (1997) Activation by cyclic GMP binding causes an apparent


2151

29.

30.

31.

32. 33.

34. 35.

36.

37. 38.

39. 40. 41.

42. 43.

2152

conformational change in cGMP-dependent protein kinase. J. Biol. Chem. 272, 31922–31928 Jarchau, T., Haeusler, C., Markert, T., Poehler, D., Vandekerckhove, J., De Jonge, H. R., Lohmann, S. M., and Walter, U. (1994) Cloning, expression, and in situ localization of rat intestinal cGMP-dependent protein kinase II. Proc. Natl. Acad. Sci. USA 91, 9426 –9430 Meinecke, M., Geiger, J., Butt, E., Sandberg, M., Jahnsen, T., Chakraborty, T., Walter, U., Jarchau, T., and Lohmann, S. (1994) Human cyclic GMP-dependent protein kinase Ib overexpression increases phosphorylation of an endogenous focal contact-associated vasodilator-stimulated phosphoprotein without altering the thrombin-evoked calcium response. Mol. Pharmacol. 46, 283–290 Vaandrager, A. B., Smolenski, A., Tilly, B. C., Houtsmuller, A. B., Ehlert, E. M. E., Bot, A. G. M., Edixhoven, M., Boomaars, W. E. M., Lohmann, S. M., and De Jonge, H. R. (1998) Membrane targeting of cGMP-dependent protein kinase is required for cystic fibrosis transmembrane conductance regulator Cl2channel activation. Proc. Natl. Acad. Sci. USA 95, 1466 – 1471 Fisch, T. M., Prywes, R., Simon, C., and Roeder, R. G. (1989) Multiple sequence elements in the c-fos promoter mediate induction by cAMP. Genes Dev. 3, 198 –211 Markert, T., Vaandrager, A. B., Gambaryan, S., Poehler, D., Haeusler, C., Walter, U., De Jonge, H. R., Jarchau, T., and Lohmann, S. M. (1995) Endogenous expression of type II cGMP-dependent protein kinase mRNA and protein in rat intestine. J. Clin. Invest. 96, 822– 830 Boerth, N. J., Dey, N. B., Cornwell, T. L., and Lincoln, T. M. (1997) Cyclic GMP-dependent protein kinase regulates vascular smooth muscle cell phenotype. J. Vasc. Res. 34, 245–259 Haby, C., Lisovoski, F., Aunis, D., and Zwiller, J. (1994) Stimulation of the cyclic GMP pathway by NO induces expression of the immediate early genes c-fos and junB in PC12 cells. J. Neurochem. 62, 496 –501 Amir, S., Rackover, M., and Funk, D. (1998) Blockers of nitric oxide synthase inhibit stress activation of c-fos expression in neurons of the hypothalamic paraventricular nucleus in rats. Neuroscience 77, 623– 627 Morris, B. J. (1995) Stimulation of immediate early gene expression in striatal neurons by nitric oxide. J. Biol. Chem. 270, 24740 –24744 Sheng, M., McFadden, G., and Greenberg, M. E. (1990) Membrane depolarization and calcium induce c-fos transcription via phosphorylation of transcription factor CREB. Neuron 4, 571– 582 Greenberg, M., and Ghosh, A. (1995) Calcium signaling in neurons: Molecular mechanisms and cellular consequences. Science 268, 239 –247 Snyder, S. H., Jaffrey, S. R., and Zakhary, R. (1998) Nitric oxide and carbon monoxide: parallel roles as neural messengers. Brain Res. Rev. 26, 167–175 Gambaryan, S., Ha¨usler, C., Markert, T., Po¨hler, D., Jarchau, T., Walter, U., Haase, W., Kurtz, A., and Lohmann, S. M. (1996) Expression of type II cGMP-dependent protein kinase in rat kidney is regulated by dehydration and correlated with renin gene expression. J. Clin. Invest. 98, 662– 670 Uhler, M. (1993) Cloning and expression of a novel cyclic GMP-dependent protein kinase from mouse brain. J. Biol. Chem. 268, 13586 –13591 Thelen, M., Rosen, A., Nairn, A. C., and Aderem, A. (1991) Regulation by phosphorylation of reversible association of a myristoylated protein kinase C substrate with the plasma membrane. Nature (London) 351, 320 –322

Vol. 13

December 1999

44.

45. 46. 47.

48.

49. 50.

51. 52.

53. 54. 55.

56.

57.

58.

59.

Po¨hler, D., Butt, E., Meissner, J., Mu¨ller, S., Lohse, M., Walter, U., Lohmann, S. M., and Jarchau, T. (1995) Expression, purification and characterization of the cGMP-dependent protein kinase Ib and II using the baculovirus system. FEBS Lett. 374, 419 – 425 Vo, N. K., Gettemy, J. M., and Coghlan, V. M. (1998) Identification of cGMP-dependent protein kinase anchoring proteins (GKAPs). Biochem. Biophys. Res. Commun. 246, 831– 835 Johnston, H. M., and Morris, B. J. (1994) NMDA and nitric oxide increase microtubule-associated protein 2 gene expression in hippocampal granule cells. J. Neurochem. 63, 379 –381 Johnston, H. M., and Morris, B. J. (1995) N-Methyl-D-aspartate and nitric oxide regulate the expression of calcium/calmodulindependent kinase II in the hippocampal dentate gyrus. Mol. Brain Res. 31, 141–150 Idriss, S. D., Gudi, T., Casteel, D. E., Kharitonov, V. G., Pilz, R. B., and, and Boss, G. R. (1999) Nitric oxide regulation of gene transcription via soluble guanylate cyclase and type I cGMP-dependent protein kinase. J. Biol. Chem. 274, 9489 –9493 Murphy, B. J., and Scott, J. D. (1998) Functional anchoring of the cAMP-dependent protein kinase. Top. Chem. Mutagen. 8, 89 –95 Snyder, G. L., Girault, J. -A., Chen, J. Y. C., Czernik, A. J., Kebabian, J. W., Nathanson, J. A., and Greengard, P. (1992) Phosphorylation of DARPP-32 and protein phosphatase inhibitor-1 in rat choroid plexus: regulation by factors other than dopamine. J. Neurosci. 12, 3071–3083 Schonthal, A., Tsukitani, Y., and Feramisco, J. R. (1991) Transcriptional and post-transcriptional regulation of c-fos expression by the tumor promoter okadaic acid. Oncogene 6, 423– 430 Alberts, A. S., Montminy, M., Shenolikar, S., and Feramisco, J. R. (1994) Expression of a peptide inhibitor of protein phosphatase 1 increases phosphorylation and activity of CREB in NIH 3T3 fibroblasts. Mol. Cell. Biol. 14, 4398 – 4407 Metz, R., and Ziff, E. (1991) cAMP stimulates the C/EBP-related transcription factor rNFIL-6 to trans-locate to the nucleus and induce c-fos transcription. Genes Dev. 5, 1754 –1766 Chinery, R., Brockman, J. A., Dransfield, D. T., and Coffey, R. J. (1997) Antioxidant-induced nuclear translocation of CCAAT/ enhancer-binding protein b. J. Biol. Chem. 227, 30356 –30361 Son, H., Hawkins, R. D., Martin, K., Kiebler, M., Huang, P. L., Fishman, M. C., and Kandel, E. R. (1996) Long-term potentiation is reduced in mice that are doubly mutant in endothelial and neuronal nitric oxide synthase. Cell 87, 1015–1023 Kleppisch, T., Pfeifer, A., Klatt, P., Ruth, P., Montkowshi, A., Fassler, R., and Hofmann, F. (1999) Long-term potentiation in the hippocampal CA1 region of mice lacking cGMP-dependent kinases is normal and susceptible to inhibition of nitric oxide synthase. Neuroscience 19, 48 –55 Blitzer, R. D., Connor, J. H., Brown, G. P., Wong, T., Shenolikar, S., Iyengar, R., and Landau, E. M. (1998) Gating of CaMKII by cAMP-regulated protein phosphatase activity during LTP. Science 280, 1940 –1943 Nakazawa, K., Karachot, L., Nakabeppu, Y., and Yamamori, T. (1993) The conjunctive stimuli that cause long-term desensibilization also predominantly induce c-Fos and Jun-B cerebellar Purkinje cells. NeuroReport 4, 1275–1278 Ohki, K., Yoshida, K., Hagiwara, M., Harada, T., Takamura, M., Ohashi, T., Matsuda, H., and Imaki, J. (1995) Nitric oxide induces c-fos gene expression via cyclic AMP element binding protein (CREB) phosphorylation in rat retinal pigment epithelium. Brain Res. 696, 140 –144

The FASEB Journal

Received for publication February 23, 1999. Revised for publication June 7, 1999.

GUDI ET AL.