An increase in the expression of the glial fibrillary acidic protein (GFAP) gene by astrocytes appears to constitute a crucial component of the brain's response to injury because it is seen in many different species and features prominently in diverse neurological diseases. Previously, we have used a modified GFAP gene (C-339) to target the expression of beta-galactosidase (beta-gal) to astrocytes in transgenic mice (Mucke et al.; New Biol 3:465-474 1991). To determine to what extent the in vivo expression of GFAP-driven fusion genes is influenced by intragenic GFAP sequences, the E. coli lacZ reporter gene was either placed downstream of approximately 2 kb of murine GFAP 5' flanking region (C-259) or ligated into exon 1 of the entire murine GFAP gene (C-445). Transgenic mice expressing C-259 versus C-445 showed similar levels and distributions of beta-gal activity in their brains. Exclusion of intragenic GFAP sequences from the GFAP-lacZ fusion gene did not diminish injury-induced upmodulation of astroglial beta-gal expression or increase beta-gal expression in non-astrocytic brain cells. These results demonstrate that 2 kb of murine GFAP 5' flanking region is sufficient to restrict transgene expression primarily to astrocytes and to mediate injury-responsiveness in vivo. This sequence therefore constitutes a critical target for mediators of reactive astrocytosis. While acute penetrating brain injuries induced focal increases in beta-gal expression around the lesion sites in C-259, C-445, and C-339 transgenic mice, infection of C-339 transgenic mice with scrapie led to a widespread upmodulation of astroglial beta-gal expression. Hence, GFAP-lacZ transgenic mice can be used to monitor differential patterns of astroglial activation in vivo. These and related models should facilitate the assessment of strategies aimed at the in vivo manipulation of GFAP expression and astroglial activation.