Transcription factor ATF2 cooperates with v-Jun to promote growth factor-independent proliferation in vitro and tumor formation in vivo.

ATF2 belongs to the bZIP family of transcription factors and controls gene expression via 8-bp ATF/CREB motifs either as a homodimer or as a heterodimer-for instance, with Jun-but has never been shown to be directly involved in oncogenesis. Experiments were designed to evaluate a possible role of ATF2 in oncogenesis in chick embryo fibroblasts (CEFs) in the presence or absence of v-Jun. We found that (i) forced expression of ATF2 cannot alone cause transformation, (ii) overexpression of ATF2 plus v-Jun specifically stimulates v-Jun-induced growth in medium with a reduced amount of serum, and (iii) the efficiency of low-serum growth correlates with the activity of a Jun-ATF2-dependent model promoter in stably transformed CEFs. Analysis of ATF2 and Jun dimerization mutants showed that the growth-stimulatory effect of ATF2 is likely to be mediated by v-Jun-ATF2 heterodimers since (i) v-Jun-m1, a mutant with enhanced affinity for ATF2, induces growth in low-serum medium much more efficiently than v-Jun, when expressed alone or in combination with ATF2; and (ii) ATF2/fos, a mutant that efficiently binds to v-Jun but is unable to form stable homodimers, shows enhanced oncogenic cooperation with v-Jun. In addition, we examined the role of ATF2 in tumor formation by subcutaneous injection of CEFs into chickens. In contrast to v-Jun, v-Jun-m1 gave rise to numerous fibrosarcomas while coexpression of ATF2 and v-Jun-m1 led to a dramatic development of fibrosarcomas visible within 1 week. Together these data demonstrate that overexpressed ATF2 potentiates the ability of v-Jun-transformed CEFs to grow in low-serum medium in vitro and contributes to the formation of tumors in vivo.

Autocrine growth and anchorage independence: two complementing Jun-controlled genetic programs of cellular transformation.

Cellular transformation can be achieved by constitutive activation of growth-regulatory signaling pathways, which, in turn, activate nuclear transcription factors thought to execute a transformation-specific program of gene expression. Members of the dimeric transcription factor family AP-1 are at the receiving end of such growth-regulating pathways and the viral form of the AP-1 subunit Jun establishes one important aspect of transformation in chick embryo fibroblasts (CEFs): enhanced growth in agar and in low serum. Enhanced Jun activity is likely to target several different genetic programs as Jun forms heterodimers with one of several members of the Fos and ATF2 subfamilies, resulting in transcription factors with different sequence specificities. To identify the programs relevant for transformation, we have reduced the complexity of AP-1 factors by constructing Jun bZip mutants that can efficiently dimerize and transactivate with only a restricted set of partner subunits. Upon introduction into CEFs, a Jun mutant selective for the Fos family induced anchorage-independent growth but no growth factor-independence. In contrast, a c-Jun mutant with preference for ATF2-like proteins caused growth factor-independence, but no growth in agar. Coexpression of both mutants reestablished the combined transformation program as induced by wild-type Jun. These data show that Jun-dependent cell transformation can be resolved into at least two distinct and independent processes, anchorage and growth factor independence, obviously triggered by two classes of Jun heterodimers likely regulating different sets of target genes.

Tumor induction by v-Jun homodimers in chickens.

To study the contribution of v-Jun homodimers to oncogenesis, we constructed artificial v-Jun derivatives in which the natural dimerization domain of v-Jun was replaced by an heterologous homodimerization domain from either the viral EB1 or the yeast GCN4 transcription factor. The resulting v-Jun chimeric proteins, called v-Juneb1 and v-Jungcn4, which can no longer dimerize with Jun or Fos, should only form homodimers in the cell. Helper-independent retroviruses expressing v-Jun, v-Juneb1 and v-Jungcn4 were generated. All three viruses transformed primary cultures of chick embryo cells with the same high efficiency and promoted local tumor growth after subcutaneous injection of infected cells in young animals. In contrast, after intravenous injection of viral suspensions into chick embryos, only the chimeric proteins produced internal tumors that were lethal. These tumors were leiomyosarcomas located within the liver and along the digestive tract. Thus, in vivo, v-Juneb1 and v-Jungcn4 are more potent oncoproteins than v-Jun. These data demonstrate that when forced to accumulate, v-Jun homodimers can induce tumors efficiently. They also show that the oncogenic potential of v-Jun can be regulated through the properties of its dimerization domain.

Increased transforming activity of JunB and JunD by introduction of an heterologous homodimerization domain.

The closely-related proteins c-Jun, JunB and JunD form a family of transcription factors which require dimerization for DNA-binding and transcriptional activity. Dimerization is mediated by a conserved amphipathic alpha-helix located adjacent to a highly charged DNA-binding domain. The Jun proteins can form both homo- and heterodimers within the Jun family and can also cross-dimerize with the Fos proteins. When expressed at high levels in primary chicken cells, each mouse Jun displays distinct transforming capacities: c-Jun transforms efficiently, JunB transforms poorly, and JunD does not transform at all. The composition of the transforming dimers, however, is unknown. To study the activity of Jun-Jun homodimers we constructed artificial derivatives, denoted Juneb1, in which the naturally occurring dimerization domain has been replaced by an heterologous homodimerization domain from the Epstein-Barr virus transcription factor EB1. These derivatives were introduced into chicken cells and assayed for their ability to affect growth. Unexpectedly, all three Juneb1 proteins conferred a transformed phenotype to primary cultures, promoting sustained growth in low-serum medium and colony formation from single cells in agar. These data demonstrate that when forced to accumulate as homodimers, both JunB and JunD can transform cells. They also suggest that the poor transforming activity of JunB and the absence of transforming activity of JunD may be due to their inability to accumulate to high levels as homodimers.