Retinoblastoma protein interacts with ATF2 and JNK/p38 in stimulating the transforming growth factor-beta2 promoter.

Two highly related transcription factors, ATF2 and ATFa, enhance the activity of the Transforming Growth Factor beta2 (TGF-beta2) promoter via a partial cAMP response element in transfected CHO cells. The retinoblastoma protein (Rb) also activates this promoter and enhances the stimulatory effects of ATF2 but causes near extinction of the effects of ATFa. The site on Rb required for its effects alone and in combination with the ATFs has been mapped mainly to the A/B pockets but the C pocket is also implicated. Whereas MKK7 or JNK expression enhances the actions of both ATFs, MKK6 or p38 expression only augments the effects of ATF2. Immunoprecipitation with Rb antibodies of lysates from transfected cells brings down expressed ATF2 but not ATFa. Expressed JNK and p38 are also found in the anti-Rb immunoprecipitates. ATF2 antibodies bring down expressed Rb, JNK and p38 and expression of Rb enhances the immunoprecipitation of both JNK and p38 by ATF2 antibodies. The results suggest that Rb is acting as a matchmaker by bridging either JNK or p38 with their common substrate ATF2 and, hence, facilitating its activation. Consistent with this suggestion, expression of Rb enhances the phosphorylation of ATF2 in CHO cells.

Interaction of retinoblastoma gene product with transcription factors ATFa and ATF2.

Two highly related transcription factors, activating transcription factors a and 2 (ATFa and ATF2) are able to activate expression in CHO cells of a reporter gene driven by the human transforming growth factor beta 2 gene promoter. This action is dependent on the intact motif CGTCA which is found as part of the cAMP responsive element in a number of promoters and to which both factors can bind in vitro. The retinoblastoma gene product also weakly stimulates expression of this reporter gene but, in combination with the factors, it exhibits a differential action: additive or greater stimulatory effects with ATF2 but strong inhibition of the actions of ATFa. Thus, although both of these two related factors are able to exert the same apparent effect on this promoter, coexpression of the retinoblastoma gene product reveals at least one significant difference in their actions.

Characterization of rat liver nuclear proteins which recognize the cAMP responsive element.

1. The data herein reveal the existence of cAMP-responsive element (CRE)-binding factors (CRF) in the nuclear extracts from cAMP-treated rat liver. 2. DNAase I and DMS footprinting analysis showed that the CRFs protected the CRE (-77 to -92) in the phosphoenolpyruvate carboxykinase (PEPCK) promoter and the TGACGTCA motif in a consensus oligodeoxynucleotide based on the sequence of the CRE's of 6 cAMP-regulated genes (C32mer). 3. Competition assays indicate that the CRF(s) is a CGTCA-specific, ATF/CREB-like factor(s). 4. Southwestern (SW) blot analysis detected 2 apparent CRFs which have molecular weights of about 30 and 32 kDa, respectively. 5. Based on the comparison of the size and binding specificity of the CRFs with the CREBs reported to date, the CRFs appear to be novel CRE-binding nuclear factors.

Interaction of a nuclear factor 1-like protein with a cAMP response element-binding protein in rat liver.

1. The existence of both cAMP-responsive element binding factor and a nuclear factor 1-like (NF-1-like) protein in nuclear extracts from liver of cAMP-treated rat has been revealed. 2. Binding of these proteins to a DNA fragment containing both elements was cooperative, and 50% binding was achieved with considerably less protein than with a fragment bearing either element alone. 3. Cleavage of the fragment between the two elements abolished the apparent cooperative interaction. 4. Southwestern blot analysis showed that the NF-1-like protein has a molecular weight in the 28-30-kDa range. 5. The NF-1-like binding activity was very stable.

Functional recognition of the neuronal tyrosine hydroxylase cAMP regulatory element in different cell types.

Transient transfection of pLB2CAT constructs bearing short synthetic oligonucleotides derived either from the tyrosine hydroxylase (TH) promoter or other sources was used to examine functional cAMP regulatory element (CRE) activity in a variety of cell lines. The region containing only the putative TH CRE was found to be as or more effective in conferring cAMP responsiveness onto pLB2CAT (which employs the TK promoter) than the immediate 272 bp region of the TH promoter. Increases in CAT activity of 10- to 20-fold were observed in JEG-3 cells with a single insert of the TH CRE region (-31 to -54) in pLB2CAT, and the presence of a second insert generated only a modest further increase. This construct also responded to cAMP in 4 other cell lines tested but the degree of increase was less dramatic. Inserts containing the consensus 8 bp CRE motif embedded in other natural or artificial contexts served generally as weak functional CREs in all cell lines tested. In vitro analysis revealed that a specific protein-DNA complex apparently containing a single protein with a MW of 45-50 kDa was formed equally well with JEG-3 cell nuclear extract and CRE-bearing-TH and other fragments which produced dramatically different cAMP effects in vivo. These results suggest specificity in the effects of cAMP on different CREs which are dictated by contextual differences.

Evidence for a cAMP-dependent nuclear factor capable of interacting with a specific region of a eukaryotic gene.

Nuclear extracts prepared from the livers of rats treated with or without 8-bromo-cAMP were tested for their ability to bind to various fragments from the flanking region of the gene encoding phosphoenolpyruvate carboxykinase (GTP) [GTP: oxaloacetate carboxy-lyase (transphosphorylating), EC 4.1.1.32] known to contain the element that confers transcriptional regulation by cAMP. Using the nitrocellulose-filtration method, concentration-dependent, apparently saturable binding was seen that is both specific and cAMP dependent. Analysis of various fragments pinpointed the active binding region to positions within the -67 to -111 region, which coincides with the functional regulatory element as shown by recent transfection studies. Formation of an apparently single complex between a synthetic oligomer containing the region from -67 to -111 of the phosphoenolpyruvate carboxykinase gene and a factor in nuclear extracts from cAMP-treated rat liver was visualized by the gel-retardation method. Complex formation is both concentration and cAMP dependent and can be prevented by excess specific but not nonspecific competitor DNA. The congruity of the results with the two different methods suggests that the factor we have detected has properties consistent with a possible role as mediator of the transcriptional control exerted by cAMP in eukaryotic cells.

Cyclic adenonosine monophosphate does not affect the stability of the messenger ribonucleic acid for tyrosine aminotransferase in cultured hepatoma cells.

Cyclic AMP has been shown to stimulate synthesis of tyrosine aminotransferase (L-tyrosine:2-oxoglutarate aminotransferase, EC 2.6.1.5) by increasing the amount of its mRNA through an increase in initiation of transcription. However, cAMP also has posttranscriptional effects on the enzyme's synthesis, as evidenced by the 4- to 5-fold enhanced decline seen when cultured hepatoma cells are exposed to cAMP and transcription is inhibited. As a direct test of the possibility that cAMP exerts this effect by destabilizing the mRNA for tyrosine aminotransferase, we analyzed the rate of decay of the mRNA using the transcriptional inhibitor 5,6-dichlororibofuranosylbenzimidazole, Northern blot analysis, and an internal standard consisting of prelabeled rRNA. It was found that the half-life of the mRNA (2.0 +/- 0.2 h) was not changed by treatment of cultured hepatoma cells under conditions which increase intracellular cAMP levels. These mRNA half-life values were not significantly different from the decline in the rate of synthesis of the enzyme after induction in dexamethasone-treated cells. We conclude that cAMP does not affect the stability of the mRNA for tyrosine aminotransferase and discuss other possible explanations for the paradoxical effect of cAMP on deinduction of this enzyme.

Progesterone decreases DNA binding factor activity and the expression in Xenopus oocytes of a cAMP responsive gene from rat liver.

Stage VI Xenopus oocytes were injected with a plasmid (pBB0.6-CAT) which contains the cAMP regulatory element (CRE) from the rat liver phosphoenolpyruvate carboxykinase (PEPCK) gene fused upstream from a reporter gene [chloramphenicol acetyltransferase (CAT)]. Inhibition of the expression of the reporter gene (average = 51%) was observed in the presence of 10 microM progesterone, which is known to lead to inactivation of the oocyte cAMP dependent protein kinase (A kinase). In contrast, oocytes injected with a control plasmid (pSV2CAT), which contains no CRE, exhibited a variable increase (average = 31%) in CAT activity after progesterone treatment. Injection of the purified bovine cardiac A kinase catalytic subunit prior to exposure of oocytes injected with pBB0.6 CAT to progesterone prevents the loss of CAT activity generated by incubation with the steroid. Gel retardation analyses with oocyte lysates and a labeled synthetic oligonucleotide fragment containing the CRE from the PEPCK gene showed the existence of a complex with the same Rf and specificity as that formed with rat liver extracts. Subsequent exposure to progesterone, however, led to a rapid and extensive decrease in this binding activity. Taken together, these results are consistent with but do not prove the hypothesis that progesterone treatment and A kinase inactivation lead to a decrease in pBB0.6 CAT expression by virtue of a decline in the binding activity of an oocyte factor(s) to the CRE of the PEPCK fragment in pBB0.6-CAT, thereby decreasing transcription of the CAT gene.

Is cyclic AMP dependent protein kinase responsible for the in vivo phosphorylation of tyrosine aminotransferase?

Undegraded tyrosine aminotransferase was purified to near homogeneity from rat liver and was confirmed to be a substrate for the beef heart cyclic AMP dependent protein kinase catalytic subunit. Specific antibody was used to quantitate the amount of phosphate incorporated into the enzyme. Phosphate incorporation was maximal at a catalytic subunit to tyrosine aminotransferase molar ratio of 7:1 using 200 microM ATP for 30 to 60 min at 30 degrees C. Phospho-peptide maps of tyrosine aminotransferase phosphorylated in vitro by the catalytic subunit were compared with those of amino-transferase immunoprecipitated from 32P labeled cells treated with and without 8-Br cAMP. Whereas the phospho-peptide maps of tyrosine aminotransferase isolated from cells treated with and without 8-Br cAMP were identical, differences were observed in the peptide map of tyrosine aminotransferase phosphorylated in vitro and in vivo. These results were taken to indicate that the catalytic subunit is not responsible for tyrosine aminotransferase phosphorylation in vivo.

In vitro phosphorylation of 4-aminobutyrate aminotransferase by cAMP dependent protein kinase.

Highly purified 4-aminobutyrate aminotransferase from pig brain is susceptible to phosphorylation by the purified cAMP-dependent protein kinase catalytic subunit. Up to 0.7 moles of phosphate from ATP-(gamma)-32P can be incorporated per mole of dimeric holoenzyme. Maximum phosphorylation was observed within about 90 minutes at 30 degrees C. Despite the extensive degree of phosphorylation observed, no kinetic property of the enzyme was perceptibly altered. Removal of cofactor had no detectable impact on the extent of phosphorylation but thermal inactivation of the enzyme increased and mild reduction with sodium borohydride decreased the phosphorylatability of the aminotransferase. It was possible to separate the enzyme into phospho and dephospho forms by the use of DEAE chromatography. Validation that the two fractions represented genuine aminotransferase was obtained by proteolytic peptide mapping. The phospho form of the enzyme was found to possess little or no aminotransferase activity while that of the dephospho form exhibited higher specific activity than the purified enzyme prior to phosphorylation. Furthermore, the dephospho form of the enzyme could not be detectably phosphorylated by reincubation with the kinase following DEAE chromatography unless it was subjected to thermal inactivation. The stoichiometry of phosphorylation of the fraction containing 32P from DEAE chromatography was approximately 1 mole/mole of dimer. These results suggest that the substrate for phosphorylation by the kinase is a form of the aminotransferase which is somehow inactivated during routine purification even when extensive precautions are taken to maximally preserve catalytic activity.

On the role of protein kinase subunits in the control of eukaryotic gene expression.

Transcriptional regulation by cAMP has been demonstrated for several eukaryotic genes; however, the identity of the protein kinase subunit involved has been a source of debate. Based on homologies with the procaryotic cAMP-binding catabolite activator protein, a recent hypothesis has invoked the regulatory protein RII as the mediator. The evidence currently available on the effects of microinjected kinase subunits suggests, however, that the catalytic subunit is the active factor. Moreover, the proposed homologies between the catabolite activator protein and RII are difficult to reconcile with its proposed mediatory role. We suggest as an alternative hypothesis that a phosphoprotein other than RII may mediate the effects of cAMP on eukaryotic gene expression.

Stimulation of tyrosine aminotransferase degradation by methylthioinosine.

Methylthioinosine (MeSno) is a purine nucleoside analog which is cytotoxic to a number of cultured cell lines including the Reuber H35 hepatoma cells used in the present studies. It has also been observed to cause a rapid profound loss of tyrosine aminotransferase activity in H35 cells well before the onset of any measurable cytotoxicity. The effect is both time and concentration dependent. MeSno does not acutely inhibit synthesis of the enzyme as evidenced by the ability of glucocorticoids or cAMP analogs to induce the enzyme to the same extent in the presence or absence of the drug. The enzyme in extracts of cells treated with the drug is essentially identical with the enzyme from extracts of control cells in terms of thermal stability, immunoprecipitability, and affinities for substrates and cofactor. Addition of MeSno to cell extracts and mixing experiments suggests that the thiopurine does not have any direct effect on enzyme activity. Immunochemical analysis of the rates of synthesis and degradation of the aminotransferase have shown that the enzyme is degraded approximately 3-4 times more rapidly in cells treated with the drug than in control cells. At the same time there is no inhibition of the rate of synthesis of the enzyme.

Direct evidence that the protein kinase catalytic subunit mediates the effects of cAMP on tyrosine aminotransferase synthesis.

The effect of purified beef heart cAMP-dependent protein kinase catalytic subunit on tyrosine aminotransferase activity in intact cultured rat H35 hepatoma cells was directly tested by micro-injection using human red blood cell ghosts as vehicles. Although the micro-injection procedure itself produced temporary fluctuations in protein synthesis and in tyrosine aminotransferase activity in H35 cells, after a recovery period of 8-12 h, these parameters returned to normal in parallel with restoration of full inducibility of the aminotransferase by both 8-Br-cAMP and dexamethasone. Eight to sixteen hours after fusion of H35 cells with unloaded ghosts, ghosts loaded with bovine serum albumin or mock-loaded with the partially purified protein kinase catalytic subunit, no significant change in the activity of the aminotransferase was detected. In contrast, fusion with ghosts loaded with the catalytic subunit at concentrations between 0.1-2 mg/ml caused reproducible 2-3-fold increases in enzyme activity. Homogeneous preparations of the catalytic subunit exhibited even greater potency as an inducer. The effect was both time- and concentration-dependent and was abolished by inactivation of the catalytic subunit with N-ethylmaleimide prior to loading. The partially purified inhibitor of protein kinase from beef heart, while not affecting basal tyrosine aminotransferase activity, selectively inhibited the ability of 8-Br-cAMP but not that of dexamethasone to stimulate the activity of this enzyme. In addition, micro-injection of the pure regulatory subunit of the kinase blocked the response of the aminotransferase to low concentrations of 8-Br-cAMP. These results provide strong support for the proposition that the catalytic subunit of protein kinase mediates the effects of cAMP on the synthesis of tyrosine aminotransferase.

Differences in rates of tyrosine aminotransferase deinduction with cyclic AMP and glucocorticoids.

As an indirect test of the possibility that cyclic AMP produces its stimulatory effects on the synthesis of tyrosine aminotransferase (L-tyrosine:2-oxoglutarate aminotransferase, EC 2.6.1.5) by stabilizing its mRNA, the kinetics of decline in the rate of synthesis of this enzyme was measured in rat hepatoma cells following inducer removal or addition of RNA synthesis inhibitors. In all cases in which cells were previously exposed to cyclic AMP, the decline in enzyme synthesis after removal of inducer or addition of inhibitors of RNA synthesis was 4 to 5 times more rapid than in cells exposed to a glucocorticoid (dexamethasone) which also induces the aminotransferase. Thus, it seems unlikely that cyclic AMP had been acting by stabilizing the mRNA that directs the synthesis of the aminotransferase. Possible explanations for these paradoxical results are suggested. A more rapid decline in enzyme synthesis was also seen in cells induced with both cyclic AMP and glucocorticoid, suggesting direct interaction between the two inducers.

Studies of cAMP metabolism in cultured hepatoma cells: presence of functional adenylate cyclase despite low cAMP content and lack of hormonal responsiveness.

The ability of isoproterenol, glucagon, PGE1 and cholera toxin to stimulate the synthesis of cAMP and protein kinase activity in line of liver cells (BRL) and a line of rat hepatoma cells (H35) has been determined. The concentration of cAMP in BRL cells (approximately 10 pmoles/mg protein) is in the range reported for other cultured cell lines but H35 cells contain extraordinarily low amounts of this cyclic nucleotide (approximately 0.05 pmoles/mg protein). Isoproterenol and PGE1 caused an increase in cAMP content, and protein kinase activation in BRL cells, although glucagon was ineffective. H35 cells, in contrast, were completely insensitive to all hormonal agonists. Despite this fact, cholera toxin was able to produce a marked increase in cAMP content, adenylate cyclase activity and protein kinase activation in H35 cells. binding studies with [125 I]-iodohydroxybenzylpindolol, a specific beta-adrenergic receptor antagonist, revealed that each H35 cell possesses fewer than 10 beta-adrenergic receptors whereas BRL cells contain 2-5,000 receptors per cell. The low level of cAMP in H35 cells appears to result from a combination of totally unstimulated adenylate cyclase and apparently elevated phosphodiesterase activities.