Search

Our previous studies have shown that the 49-base pair region of promoter DNA between -1719 and -1670 base pairs is necessary for transcription of the rat COL1A1 gene in transgenic mouse calvariae. In this study, we further define this element to the 13-base pair region between -1683 and -1670. This element contains a TAAT motif that binds homeodomain-containing proteins. Site-directed mutagenesis of this element in the context of a COL1A1-chloramphenicol acetyltransferase construct extending to -3518 base pairs decreased the ratio of reporter gene activity in calvariae to tendon from 3:1 to 1:1, suggesting a preferential effect on activity in calvariae. Moreover, chloramphenicol acetyltransferase-specific immunofluorescence microscopy of transgenic calvariae showed that the mutation preferentially reduced levels of chloramphenicol acetyltransferase protein in differentiated osteoblasts. Gel mobility shift assays demonstrate that differentiated osteoblasts contain a nuclear factor that binds to this site. This binding activity is not present in undifferentiated osteoblasts. We show that Msx2, a homeodomain protein, binds to this motif; however, Northern blot analysis revealed that Msx2 mRNA is present in undifferentiated bone cells but not in fully differentiated osteoblasts. In addition, cotransfection studies in ROS 17/2.8 osteosarcoma cells using an Msx2 expression vector showed that Msx2 inhibits a COL1A1 promoter-chloramphenicol acetyltransferase construct. Our results suggest that high COL1A1 expression in bone is mediated by a protein that is induced during osteoblast differentiation. This protein may contain a homeodomain; however, it is distinct from homeodomain proteins reported previously to be present in bone.

We have previously reported that the expression of the ColCAT3.6 transgene containing 3.5 kilobases (kb) of alpha 1(I) collagen (COL1A1) promoter sequence fused to the chloramphenicol acetyltransferase (CAT) reporter gene paralleled the expression of the endogenous gene in several connective tissues. We report here that the activity of the reporter gene in aorta from 7-day-old transgenic mice is 10-64-fold lower than in tendon or bone, whereas the endogenous gene is highly expressed in all three tissues. In contrast, the COL1A1 minigene containing 2.3 kb of upstream sequence, the first five exon/intron units, the last six exon/intron units, and 2 kb of 3'-flanking sequence showed high CAT activity in aorta. These results suggest that cis sequences found in ColCAT3.6 mediate high levels of COL1A1 expression in bone and tendon, but not in vascular smooth muscle cells (VSMC), whereas sequences located within the minigene, but not found in ColCAT3.6, mediate VSMC-specific expression. Analysis of promoter activity in cultured cells derived from transgenic tissues further suggests the presence of VSMC-specific regulatory domains. Transient transfection studies, however, failed to shows differential regulation. These differences stress the importance of not relying exclusively on transient transfection data when mapping tissue-specific regulatory domains.

We examined the effect of basic fibroblast growth factor (bFGF) on alpha 1(I) procollagen mRNA levels, alpha 1(I) collagen gene transcription, and alpha 1(I) collagen promoter activity in osteoblastic MC3T3-E1 cells. Cells were stably transfected with ColCAT 3.6, containing 3521 base pairs of alpha 1(I) collagen promoter DNA, fused to the CAT reporter gene, or an upstream deletion mutant of ColCAT 3.6 designated ColCAT 2.3. After 48 h, bFGF (0.1-10 nM) inhibited the incorporation of [3H]proline into collagenase-digestible protein (CDP). Indomethacin did not alter the inhibitory effect of bFGF on CDP labeling. Aphidicolin, an inhibitor of DNA synthesis, did not block the inhibitory effect of bFGF on CDP. bFGF (1-10 nM) decreased alpha 1(I) procollagen mRNA levels, with maximal inhibition, nearly 99% of control, caused by 10 nM bFGF. After 48 h, bFGF (1 nM) reduced alpha 1(I) procollagen gene transcription by about 92%. ColCAT 3.6 activity was inhibited with 0.1-10 nM bFGF and was maximally repressed by about 83% with 10 nM bFGF. In contrast, bFGF (1 and 10 nM) caused a stimulation of ColCAT 2.3 activity. These data show that bFGF inhibits collagen synthesis by a transcriptional mechanism and the alpha 1(I) collagen promoter contains DNA sequences which mediate bFGF inhibition of type I collagen gene expression in bone.

1,25-Dihydroxyvitamin D3 (1,25-(OH)2D3) inhibited type I collagen synthesis in the central bone but not the periosteum of fetal rat calvaria maintained in organ culture. The central bone synthesized primarily type I collagen, whereas the periosteum synthesized both types I and III collagen. Enhanced degradation of newly synthesized collagen seemed an unlikely mechanism for the observed decrease in collagen synthesis since 1,25-(OH)2D3 reduced both the labeling of intracellular procollagen and total [3H]hydroxyproline formation in fetal rat calvaria. As measured by cell-free translation of RNA extracted from calvaria, 1,25-(OH)2D3 decreased the level of functional procollagen mRNA and collagen synthesis to the same extent. Both collagen synthesis and procollagen mRNA levels were decreased as early as 3 h after exposure to 1,25-(OH)2D3 and were 50% of the control level by 24 hr. Upon removal of 1,25-(OH)2D3 from the cultures, collagen synthesis and procollagen mRNA remained depressed for 24 h but returned to control levels by 48 h. A reduction in collagen synthesis and procollagen mRNA was also observed in calvaria excised from 6-day-old rat pups given a single subcutaneous injection of 1,25-(OH)2D3 (1.6 ng/g body weight). We conclude that 1,25-(OH)2D3 inhibits the synthesis of type I collagen in the differentiated osteoblast by reducing the level of functional procollagen mRNA.

Our previous work demonstrated that the inhibition of type I collagen synthesis by 1,25-dihydroxyvitamin D (1,25-(OH)2D3) in fetal rat calvaria and cultured rat osteosarcoma cells is accompanied by equivalent reduction in steady state levels of alpha 1(I) and alpha 2(I) collagen mRNA. To pursue the mechanism for this effect, we isolated and sequenced a 3.6-kilobase DNA fragment that contained the promoter for the rat alpha 1(I) collagen gene. This promoter fragment was fused to the chloramphenicol acetyltransferase gene and was introduced into ROS 17/2.8 cells by calcium phosphate co-precipitation. Expression of this construct was diminished by 1,25-(OH)2D3 to the same degree as the endogenous collagen gene in both transient expression assays and in permanently selected bone cells. However, a fibroblast cell line did not show a similar reduction in the activity of the transgene or the endogenous collagen gene. These experiments indicate that the alpha 1(I) promoter contains cis-active elements which are regulated by the 1,25-(OH)2D3 receptor in ROS 17/2.8 cells.

Cytosol prepared from intestinal homogenates of vitamin D3-deficient rats demonstrate a 3.2 S protein having high affinity and low capacity for 1,25-dihydroxyvitamin D3. Appearance of the protein is dependent upon the presence of 0.3 M potassium chloride and dithiothreitol in the homogenization buffer. A 6 S protein having greater affinity for 25-hydroxyvitamin D3 than 1,25-dihydroxyvitamin D3 is observed using all conditions and is increased by the addition of rat serum to cytosol. The 3.2 S protein is neither a serum contaminant nor a component of the 6 S protein and is probably not a consituent of tissues which which are not responsive to vitamin D3.