THE MOUSE TYPE-III PROCOLLAGEN-ENCODING GENE - GENOMIC CLONING AND COMPLETE DNA-SEQUENCE

Overlapping cosmid clones were isolated that covered the entire mouse type-III collagen-encoding gene (mCol3) locus including flanking seque nces approximately 40 kb upstream and 20 kb downstream from the gene. This gene was characterized initially by restriction mapping and then followed by sequencing of 43.6 kb, including 5 kb upstream from the tr anscription start point (tsp) and all exons and introns of the entire gene. The optimal parameters for sequencing a gene of this size were d etermined by sequencing 5-10-kb fragments at different ratios of rando m and directed sequencing, and comparing their efficiency. Based on ou r experience for sequencing mCol3, we have estimated that the most cos t-efficient method was to achieve a twofold redundancy in sequencing b y using random DNA subclones as templates for sequencing prior to init iating directed DNA sequencing to close the gaps between contiguous re gions. mCol3 spans 37.6 kb from the tsp to the single polyadenylation site and contains 51 exons. The overall structure of mCol3 is similar to that of other members of the fibrillar collagen-encoding gene famil y. Several repetitive elements were located within the gene boundaries . Based on the nucleotide (nt) sequence, the predicted sizes of the mo use type-III collagen (mCOL3) mRNA and polypeptide are 4767 nt and 146 4 amino acids (aa), respectively. A comparison of mCOL3 versus the hum an type-III collagen (hCOL3) showed 91% identity at the aa level.