S-adenosylmethionine and its metabolite induce apoptosis in HepG2 cells: Role of protein phosphatase 1 and Bcl-x(S).

S-adenosylmethionine (SAMe) and its metabolite 5'-methylthioadenosine (MTA) are proapoptotic in HepG2 cells. In microarray studies, we found SAMe treatment induced Bcl-x expression. Bcl-x is alternatively spliced to produce two distinct mRNAs and proteins, Bcl-x(L) and Bcl-x(S). Bcl-x(L) is antiapoptotic, while Bcl-x(S) is proapoptotic. In this study we showed that SAMe and MTA selectively induced Bcl-x(S) in a time- and dose-dependent manner in HepG2 cells. There are three transcription start sites in the human Bcl-x gene which yield only Bcl-x(L) in control HepG2 cells. SAMe and MTA treatment did not affect promoter usage, but while one promoter yielded only Bcl-x(L), the other two yielded both Bcl-x(L) and Bcl-x(S), with Bcl-x(S) as the predominant messenger RNA (mRNA) species. Trichostatin A, 3-deaza-adenosine, cycloleucine, and okadaic acid had no effect on Bcl-x(S) induction by SAMe or MTA. Calyculin A and tautomycin, on the other hand, blocked SAMe and MTA-mediated Bcl-x(S) induction and apoptosis in a dose-dependent manner. SAMe and MTA increased protein phosphatase 1 (PP1) catalytic subunit mRNA and protein levels and dephosphorylation of serine-arginine proteins, with the latter blocked by calyculin A. The effects of SAMe and MTA on Bcl-x(S), PP1 expression, and apoptosis were also seen in 293 cells, but not in primary hepatocytes. Induction of Bcl-x(S) by ceramide in HepG2 cells also resulted in apoptosis. In conclusion, we have uncovered a highly novel action of SAMe and MTA, namely the ability to affect the cellular phosphorylation state and alternative splicing of genes, in this case resulting in the induction of Bcl-x(S) leading to apoptosis.

Inhibition of nuclear protein phosphatase activity in mouse hepatocytes by the cyanobacterial toxin microcystin-LR.

Microcystin-LR (MCLR) is a cyanobacterial hepatotoxin and protein phosphatase inhibitor that contaminates water reservoirs worldwide. MCLR localizes to the cytosol of hepatocytes, however, immunohistochemical studies indicate that it accumulates in the nucleus. MCLR toxicosis is associated with decreased hepatic protein phosphatase activity, but effects in nuclear protein phosphatase activity have not been investigated. Balb/c mice were given lethal (100 microg/kg) or sublethal (12, 23 and 45 microg/kg) i.p. doses of MCLR and hepatic nuclear extracts were analyzed for protein phosphatase 1 and 2A activity. There was profound inhibition of nuclear protein phosphatase activity within 50 min of lethal dosing, however an inhibition was not detected with sublethal doses. MCLR immunohistochemistry revealed widespread lobular staining in the lethal group and centrilobular staining in the sublethal groups. At the cellular level there was nuclear and cytoplasmic staining of equal intensity. As an indicator of nuclear protein phosphatase activity, the phosphorylation of p53, a nuclear phosphoprotein and known substrate for protein phosphatases 1 and 2A, was evaluated. Balb/c mice were treated with sublethal doses of MCLR or saline vehicle after induction of hepatic p53 by the DNA damaging agent diethylnitrosamine (DEN). P53 was immunoprecipitated and probed with phosphoserine specific antibodies by Western blotting. There was greater phosphoserine reactivity of p53 protein in animals treated with MCLR relative to saline treated controls, consistent with increased phosphorylation of serine sites. It is concluded that an interaction of this toxin with nuclear protein phosphatases occurs within 50 min of lethal dosing, which leads to a profound inhibition of enzymatic activity. Even sublethal doses of MCLR that do not result in significant inhibition of activity in bulk nuclei, result in detectable changes in phosphorylation of p53.

In vitro hepatotoxicity of the cyanobacterial alkaloid cylindrospermopsin and related synthetic analogues.

Cylindrospermopsin (CY), a sulfate ester of a tricyclic guanidine substituted with a hydroxymethyluracil, is a cyanobacterial toxin of increasing environmental import as it frequently occurs in drinking water reservoirs. As a toxin, CY mainly targets the liver but also involves other organs. In hepatocytes CY inhibits the synthesis of protein and of glutathione, leading to cell death. The total chemical synthesis of CY has recently been reported (Xie et al., 2000, J. Am. Chem. Soc. 22, 5017-5024). The synthesis has provided analogues of CY to study aspects of the relationship between chemical structure and activity that contribute to toxicity. Protein synthesis inhibition was measured in vitro using a rabbit reticulocyte system. Primary cultures of rat hepatocytes were used to determine the biological activity of CY and analogues in intact cells. Protein synthesis and cell glutathione levels were measured. We could distinguish between CY transport and biological activity by comparing the results in vitro to those in intact cells. The role of the sulfate group in CY toxicity was examined by comparing biological effects of CY with that of CY-DIOL (synthetic CY lacking the sulfate group). The sulfate group was found not to play a role in CY activity or in its uptake into cells, since there was no significant difference in biological activity in vitro or in cells between natural CY and CY-DIOL. The orientation of the hydroxyl group at C7 also had no impact on biological activity or transport of CY, since the C7 epimer of CY (EPI-CY) and the corresponding diol (EPI-DIOL) had activity similar to RAC-CY in vitro and in intact cells. AB-MODEL, the analogue lacking an intact C ring, and the methyl and hydroxyl groups of ring A could inhibit protein synthesis (but at concentrations 500-1000-fold higher than natural CY). Other structurally simpler synthetic analogues lacked biological activity.