In a 3.8-kilobase mouse DNA sequence encoding amino acid sequences for the proal(I) chain of type I procollagen, 14 coding sequences were identified which specify a sequence 95% homologous to amino acid residues 568 to %3 of the bovine al(I) chain. All of these coding sequences were flanked by appropriate splice junctions following the GT/AG rule. These observations suggest, but do not prove, that this proal(I) gene is transcriptionally active. Of the 14 coding sequences, 7 were 54 base pairs in length, whereas the remainder were higher multiples of 54 base pairs. Nonrandom utilization of codons pertained throughout all of the coding sequences showing a preference (56%) for U in the wobble position. Two of the intervening sequences encoded imperfect vestiges of coding sequences which exhibited a codon preference different from that of the proal(I) gene proper and were not flanked by splice junctions. One intervening sequence encoded a member of the mouse Bi family of middle repetitive sequences. It was flanked by 8-base-pair direct repeats and had a truncated A-rich region, suggesting that it may be a mobile element. Within this element were sequences which could function as a RNA polymerase III split promoter.

Chick cranial-bone procollagen, extracted at neutral pH in the presence of inhibitors of proteolytic enzymes, exists as a triple-stranded protein with disulfide bonds linking all three chains. The biosynthetic precursor function of this procollagen was demonstrated by pulsechase experiments. The ratio of radioactive hydroxyproline to proline in proa chains obtained by reduction and alkylation of the disulfide-bonded precursor was similar to the value determined for proal from acid-extracted procollagen. However, the molecular weight of these chains was higher than that previously determined for proal and proa2, suggesting that extraction of tissue at low pH results in a selective loss of disulfide-bonded regions from procollagen. These findings reconcile several apparently conflicting reports on the nature of procollagen identified in bone and in the medium of cultured fibroblasts and indicate that in both systems procollagen is synthesized as a disulfidelinked protein.There is ample evidence for the existence of a higher molecular weight precursor of the functional collagen monomer (1-5). The term procollagen was applied to the form identified in extracts of embryonic cranial bone to indicate an antecedent role for the molecule in the biosynthesis of collagen (1). Molecular weights of 115,000-120,000 were determined for the constituent proal and proa2 chains of cranial-bone procollagen, in contrast to the molecular weight of 95,000 established for vertebrate collagen-a chains (1, 6, 7). Possible functions for the additional sequences in procollagen include initiation of triple-helix formation, prevention of intracellular fibrogenesis, and intracellular translocation and secretion of the protein (1, 2, 8). The amino-acid composition of the additional sequence in the proal chain of chick-bone procollagen is markedly different from that of fibrous collagen, in keeping with the possibility of a distinct structure and function for this molecular domain (7,9).In contrast to the discrete proa chains obtained by heat denaturation of acid-extracted bone procollagen, collagen precursors identified in the medium of cultured cells contain interchain disulfide bonds. Thus, reduction of culture medium with 2-mercaptoethanol was required to demonstrate the presence of proa chains in the material synthesized by embryonic chick fibroblasts (3, 10). In the absence of reduction, collagen fractions with molecular weights expected of dimers and trimers of proa chains were identified (3,4,(10)(11)(12).The ability to demonstrate a form of procollagen in aceticacid extracts of cranial bone resulted largely from inactivation Abbreviations: Hepes, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; EDTA, ethylenediamine tetraacetic acid; CMcellulose, carboxymethyl cellulose. * To whom requests for reprints may be addressed. 3521at low pH of the neutral proteolytic activity, procollagen peptidase (1, 13-15). It seemed possible, however, that homogenization and extraction of bone in acetic acid might promote limited proteolysis o...

No abstract

We report the first isolation and identification of a mouse genomic fragment encoding amino acid sequences for the pro alpha 1(I) chain of type I procollagen. The DNA sequence of eight coding sequences is presented; five of these are 54 bp and three are 108 bp in length. Together these specify 198 amino acids which are 94% homologous to the corresponding bovine pro alpha 1(I) chain protein sequences. Each of the eight coding sequences is flanked by appropriate splice-junction sequences that exhibit considerable sequence complementarity to the rat small nuclear U1a RNA. In the 198 codons examined in this mouse genomic clone, the preferred codons for glycine and alanine are GGU (46/67) and GCU (23/30), respectively. This is in contrast to the codon usage reported for the chicken pro alpha 1(I) cDNA clone (Fuller and Boedtker, 1981). The examined coding sequences exhibit considerable nucleotide homology in both end-to-end and in staggered alignments. Based on an analysis of this homology data, a model is presented for the generation of 108-bp coding sequences from 54-bp units by two successive homologous recombinational events within coding sequences. Alternatively, the 108-bp units may have arisen by precise deletions of an intervening sequence between 54-bp coding sequences. Evidence supporting this is provided by a comparison of pro alpha 1(I) and pro alpha 2(I) genes. In the mouse pro alpha 1(I) gene amino acids 856-891 are encoded in a 108-bp unit; the chicken pro alpha 2(I) gene these residues are encoded in two 54-bp coding sequences. In addition, the coding sequences for nearly 50% of the alpha domain are condensed in the pro alpha 1(I) gene into a region approximately one half the size occupied by the comparable sequences in the pro alpha 2(I) gene.