Tissue-specific expression of the mouse α2(I) collagen promoter: Studies in transgenic mice and in tissue culture cells

We sought to determine the cis- acting elements responsible for the pattern of tissue specific expression of the mouse α2(I) collagen gene. Using an RNase protection assay we first verified that expression of the α2(I) collagen gene is mainly confined to tendons, bone, and skin in mice. Both transgenic mice and DNA transfection of tissue culture cells were used as experimental approaches. Transgenic mice lines were generated harboring chloramphenicol acetyltransferase (CAT) chimeric genes that contained either (a) 2000 base pairs (bp) of 5′-flanking sequences of the mouse α2(I) collagen gene plus additional sequences between +418 and +1524 of the first intron of this gene or (b) the same promoter sequences without intron sequences or (c) the 350-bp proximal promoter sequences. Transgenic mice containing both types of 2000-bp promoters showed a pattern of CAT expression that was tissue specific. The presence of sequences of the first intron in the transgene did not increase the level of promoter activity. Transgenic mice harboring the 350-bp α2(I) collagen promoter also showed a pattern that was tissue-specific except that high level expression also occurred in the brain. This suggests that negative regulation is an important component of tissue-specific expression. In order to analyze the first 350 bases in detail, we performed transient expression experiments, using promoter fragments attached to the luciferase reporter gene. Fibroblasts, which show a high level expression of the endogenous α2(I) collagen gene, and B cells, in which the gene is silent, were transfected with a series of deletions and substitution mutations within the proximal 350-bp promotor. These experiments were unable to define unique cell-specific cis-acting elements. However, when the sequence between -315 and -284 was tandemly repeated upstream of a minimal α2(I) collagen promoter (-41 to +54), the activity of this construction was considerably higher in fibroblasts than in B cells when compared with the minimal promoter itself. In gel retardation assays, the levels of complexes that bind to this sequence were higher in fibroblast nuclear extracts than in myeloma nuclear extracts. Our results are consistent with the hypothesis that the -315 to -284 DNA sequence participates in the cell-specific control of the α2(I) collagen gene in fibroblasts.