Bcl-2 oncogene expression plays a role in the establishment of persistent viral infection by blocking virus-induced apoptosis. This might be achieved by preventing virus-induced activation of caspase-3, an IL-1beta-converting enzyme (ICE)-like cysteine protease that has been implicated in the death effector phase of apoptosis. Contrary to this model, we show that three cell types highly overexpressing functional Bcl-2 displayed caspase-3 activation and underwent apoptosis in response to infection with alphaviruses Semliki Forest and Sindbis as efficiently as vector control counterparts. In all three cell types, overexpressed 26 kDa Bcl-2 was cleaved into a 23 kDa protein. Antibody epitope mapping revealed that cleavage occurred at one or two target sites for caspases within the amino acid region YEWD31 (downward arrow) AGD34 (downward arrow) A, removing the N-terminal BH4 region known to be essential for the death-protective activity of Bcl-2. Preincubation of cells with the caspase inhibitor Z-VAD prevented Bcl-2 cleavage and partially restored the protective activity of Bcl-2 against virus-induced apoptosis. Moreover, a murine Bcl-2 mutant having Asp31, Asp34 and Asp36 substituted by Glu was resistant to proteolytic cleavage and abrogated apoptosis following virus infection. These findings indicate that alphaviruses can trigger a caspase-mediated inactivation of Bcl-2 in order to evade the death protection imposed by this survival factor.

Activation of genes by heavy metals, notably zinc, cadmium and copper, depends on MTF-1, a unique zinc finger transcription factor conserved from insects to human. Knockout of MTF-1 in the mouse results in embryonic lethality due to liver decay, while knockout of its best characterized target genes, the stress-inducible metallothionein genes I and II, is viable, suggesting additional target genes of MTF-1. Here we report on a multi-pronged search for potential target genes of MTF-1, including microarray screening, SABRE selective amplification, a computer search for MREs (DNA-binding sites of MTF-1) and transfection of reporter genes driven by candidate gene promoters. Some new candidate target genes emerged, including those encoding alpha-fetoprotein, the liver-enriched transcription factor C/EBPalpha and tear lipocalin/von Ebner's gland protein, all of which have a role in toxicity/the cell stress response. In contrast, expression of other cell stress-associated genes, such as those for superoxide dismutases, thioredoxin and heat shock proteins, do not appear to be affected by loss of MTF-1. Our experiments have also exposed some problems with target gene searches. First, finding the optimal time window for detecting MTF-1 target genes in a lethal phenotype of rapid liver decay proved problematical: 12.5-day-old mouse embryos (stage E12.5) yielded hardly any differentially expressed genes, whereas at stage 13.0 reduced expression of secretory liver proteins probably reflected the onset of liver decay, i.e. a secondary effect. Likewise, up-regulation of some proliferation-associated genes may also just reflect responses to the concomitant loss of hepatocytes. Another sobering finding concerns gamma-glutamylcysteine synthetase(hc) (gamma-GCS(hc)), which controls synthesis of the antioxidant glutathione and which was previously suggested to be a target gene contributing to the lethal phenotype in MTF-1 knockout mice. gamma-GCS(hc) mRNA is reduced at the onset of liver decay but MTF-1 null mutant embryos manage to maintain a very high glutathione level until shortly before that stage, perhaps in an attempt to compensate for low expression of metallothioneins, which also have a role as antioxidants.

MTF-1 is a zinc finger transcription factor that mediates the cellular response to heavy metal stress; its targeted disruption in the mouse leads to liver decay and embryonic lethality at day E14. Recently, we have sequenced the entire MTF-1 gene in the compact genome of the pufferfish Fugu rubripes. Here we have defined the promoter sequences of human and mouse MTF-1 and the genomic structure of the mouse MTF-1 locus. The transcription unit of MTF-1 spans 42 kb (compared to 8.5 kb in Fugu) and is located downstream of the gene for a phosphatase (INPP5P) in mouse, human, and fish. In all of these species, the MTF promoter region has the features of a CpG island. In both mouse and human, the 5' untranslated region harbors conserved short reading frames of unknown function. RNA mapping experiments revealed that in these two species, MTF-1 mRNA is transcribed from a cluster of multiple initiation sites from a TATA-less promoter without metal-responsive elements. Transcription from endogenous and transfected MTF-1 promoters was not affected by heavy metal load or other stressors, in support of the notion that MTF-1 activity is regulated at the posttranscriptional level. Tissue Northern blots normalized for poly A+ RNA indicate that MTF-1 is expressed at similar levels in all tissues, except in the testes, that contain more than 10-fold higher mRNA levels.

The pufferfish Fugu rubripes was recently introduced as a new model organism for genomic studies, since it contains a full set of vertebrate genes but only 13% as much DNA as a mammal. Fugu genes tend to be smaller and densely spaced due to shortening of introns and intergenic spacers. We isolated the Fugu gene for the metal-responsive transcription factor MTF-1 (MTF1), a mediator of heavy metal regulation and oxidative stress response previously characterized in mammals. In addition, most of the cDNA sequence was also determined. The 780 amino acid MTF-1 protein of Fugu is very similar to that of mouse and human, with 90% amino acid identity in the DNA binding zinc finger domain and 57% overall identity. Expression of the pufferfish cDNA in mammalian cells shows that Fugu MTF-1 has the same DNA binding specificity as its mammalian counterpart and also induces transcription in response to zinc and cadmium. The proteincoding part of the Fugu MTF-1 gene spans 6.4 kb and consists of 11 exons. Upstream region and first exon constitute a CpG island. The distance between stop codon and polyadenylation motifs is Ͼ 2 kb, suggesting a very long 3 untranslated mRNA region, followed by another CpG island which may represent the promoter of the next gene downstream. Part of the MTF-1 genomic structure was also determined in the mouse, and some striking similarities were found: for example, the upstream adjacent gene in both species is INPP5P, encoding a phosphatase. The mouse MTF-1 promoter is also embedded in a CpG island, which however shares no sequence similarity to the one of Fugu. The Fugu CpG island is shorter than the one of the mouse and has no elevated [G+C] content; these and other data indicate that CpG islands of fish may represent a primordial stage of CpG island evolution.