The dimeric transcription factor AP-1 (activator protein-1) is composed of members of the Fos, Jun, and ATF family of proteins (Angel and Karin 1991). AP-1 activity is induced by a vast number of stimuli, including growth factors, cytokines, and UV irradiation, and couples cell surface signals to changes in cellular phenotype by regulating the expression of target genes (Angel and Karin 1991; Karin et al. 1997). Members of the AP-1 complex are critically involved in a multitude of cellular processes such as proliferation, differentiation, cell death, and oncogenic transformation. It has been hypothesized that the ability to regulate such a large number of biological processes is achieved by the formation of various functional dimers with distinct biological properties. Fos proteins (c-Fos, FosB, Fra-1, Fra-2) form stable dimers with Jun proteins (c-Jun, JunB, JunD), whereas Jun proteins can form homodimers and heterodimers with Fos and ATF proteins (Angel and Karin 1991; Karin et al. 1997). c-Fos (encoded by Fos) and Fra-1 (Fos-related antigen-1, encoded by the Fos-like-1 gene, Fosl1) show high homology in their leucine zipper and DNA-binding domains. Consistent with this sequence conservation, DNA binding specificity of Fra-1/c-Jun heterodimers is indistinguishable from c-Fos/c-Jun heterodimers (Cohen et al. 1989). Outside these domains, the proteins show little similarity. Most importantly, Fra-1 lacks transactivation domains and fails to activate transcription when fused to the DNA-binding domain of Gal4. In contrast, c-Fos is a potent activator of transcription (Cohen et al. 1989). Interestingly, Fosl1 is a transcriptional target of c-Fos, and induction of Fosl1 upon growth factor stimulation is delayed compared with Fos (Cohen and Curran 1988; Bergers et al. 1995; Schreiber et al. 1997; Matsuo et al. 2000). This led to the hypothesis that Fra-1 might counteract the function of c-Fos by lowering the transactivation potential of AP-1 complexes (Yoshioka et al. 1995). Fosl1 and Fos show different yet partially overlapping expression patterns in mice. Fosl1 is expressed in extraembryonic tissues during mouse development and in brain, skin, and testes of adult mice (Schreiber et al. 1997, 2000). Fos is expressed in the placenta and in several organs of the embryo, including fetal liver, heart, brain, and developing bones. In adult mice, Fos is readily detectable in most tissues, with highest levels found in brain and bone (Smeyne et al. 1993). Gain-of-function experiments have shown that overexpression of Fra-1 leads to increased bone formation and osteosclerosis (Jochum et al. 2000), whereas overexpression of c-Fos results in osteo- and chondrosarcomas (Ruther et al. 1987; Wang et al. 1992). Targeted disruption of the Fosl1 gene leads to early embryonic lethality due to impaired vascularization of the placenta (Schreiber et al. 2000). In contrast, lack of c-Fos results in pleiotropic defects, most importantly the bone disease osteopetrosis, due to the lack of bone-resorbing osteoclasts (Johnson et al. 1992; Wang et al. 1992; Grigoriadis et al. 1994). Furthermore, c-Fos-deficient photoreceptor neurons in the retina are resistant to light-induced apoptosis (Hafezi et al. 1997; Wenzel et al. 2000). These results suggest that c-Fos and Fra-1 have distinct functions in mice. On the other hand, Fosl1 is a transcriptional target of c-Fos in osteoclast progenitors, and overexpression of Fra-1 by viral gene transfer in vitro or in transgenic mice was able to rescue c-Fos-dependent osteoclast differentiation (Matsuo et al. 2000). To test whether Fra-1 can take over c-Fos functions under physiological conditions, knock-in mice were generated in which Fos is deleted and replaced by Fosl1. Interestingly, Fosl1 can fully complement for the lack of c-Fos in bone development and substitutes for the pro-apoptotic function of c-Fos in photoreceptor neurons. Rescue of bone cell differentiation, but not photoreceptor apoptosis, is gene dosage dependent. Furthermore, Fra-1 fails to rescue the expression of c-Fos target genes in fibroblasts. These data suggest that transactivation by c-Fos might be dispensable in vivo and that stage- and cell-specific modulation of expression determines the specific functions of these genes.