Human alpha 1 (XIII) collagen gene. Multiple forms of the gene transcripts are generated through complex alternative splicing of several short exons.

The structure of the gene for the human alpha 1(XIII) collagen chain (COL13A1) was determined from genomic clones spanning 140,000 base pairs (bp), including about 3,000 bp of the 5'-end-flanking region and 5,000 bp of the 3'-end-flanking region. The gene was shown to contain 39 exons. There were eight exons of 27 bp, five of 36 bp, four of 54 bp, three of 45 bp, and two of 42 bp. The rest of the exons coding for translated sequences had sizes varying between 24 and 153 bp. The genomic clones did not contain exons 3 and 4 whose sizes could, however, be estimated from cDNA clones. S1 nuclease mapping and primer extension analyzes indicated five closely spaced initiation sites of transcription. Sequencing of the 5'-end-flanking region did not reveal a typical TATA box but a four times repeated TATTTAT sequence that may serve as true TATA boxes. Two CCAAT boxes were found starting at positions-13 and -194, and furthermore, the promoter region contains two GC boxes. Previous studies on alpha 1 (XIII) collagen cDNA and genomic clones showed that the primary transcript undergoes complex alternative splicing generating at least four different forms of mRNAs. The present work demonstrated that sequences of seven exons are alternatively used. These exons contain sequences coding for pure collagenous regions, pure noncollagenous regions, and an exon coding for a junction of a collagenous and noncollagenous domain.

Assignment of the human collagen alpha 1 (XIII) chain gene (COL13A1) to the q22 region of chromosome 10.

Type XIII collagen is a recently described collagen that resembles in structure the short-chain collagens of types IX, X, and XII. Unlike any other collagen, the type XIII is found in several different forms generated through alternative splicing. A 2.0-kb genomic fragment from the human alpha 1 (XIII) collagen gene was isolated and shown by DNA sequencing to contain exon 12 as counted from the 3' end. This fragment was used as a probe to localize the gene. The gene (COL13A1) was assigned to chromosome 10 by hybridization of the probe to DNA isolated from a panel of human-mouse somatic cell hybrids containing different human chromosomes. Furthermore, the gene was mapped to the q22 region by in situ hybridization to metaphase chromosomes.

Gene structure for the alpha 1 chain of a human short-chain collagen (type XIII) with alternatively spliced transcripts and translation termination codon at the 5' end of the last exon.

Two overlapping human genomic clones that encode a short-chain collagen, designated alpha 1(XIII), were isolated by using recently described cDNA clones. Characterization of the cosmid clones that span approximately equal to 65,000 base pairs (bp) of the 3' end of the gene established several unusual features of this collagen gene. The last exon encodes solely the 3' untranslated region and it begins with a complete stop codon. The 10 adjacent exons vary in size from 27 to 87 bp and two of them are 54 bp. Therefore, the alpha 1-chain gene of type XIII collagen has some features found in genes for fibrillar collagens but other features that are distinctly different. Previous analysis of overlapping cDNA clones and nuclease S1 mapping of mRNAs indicated one alternative splicing site causing a deletion of 36 bp from the mature mRNA. The present study showed that the 36 bp is contained within the gene as a single exon and also that the gene has a 45-bp -Gly-Xaa-Xaa- repeat coding exon not found in the cDNA clones previously characterized. Nuclease S1 mapping experiments indicated that this 45-bp exon is found in normal human skin fibroblast mRNAs. Accordingly, the data demonstrate that there is alternative splicing of at least two exons of the type alpha 1(XIII)-chain gene.

Large introns in the 3' end of the gene for the pro alpha 1 (IV) chain of human basement membrane collagen.

Using a recently characterized cDNA clone (HT-21) coding for the pro alpha 1 (IV) chain of human type IV procollagen, we have isolated three clones from a bacterio-phage lambda Charon 4A library of human genomic DNA. The intron/exon structure of the pro alpha 1 (IV) genomic clones was analyzed by heteroduplex electron microscopy and nucleotide sequencing. The analysis showed that the introns separating exons 2-9 are large and have a total length of over 12,000 base pairs (bp). Six of seven exons at the 3' end of the gene coded for -Gly-Xaa-Yaa-repeats of the collagenous part of the chain. Five of the -Gly-Xaa-Yaa- coding exons (numbers 5-9) varied in size between 72 bp and 134 bp, and none of them were 54 bp or multiples thereof. A sixth exon (exon 4) was a junction exon containing 71 bp coding for -Gly-Xaa-Yaa- sequences and 142 bp coding for the carboxyl-terminal noncollagenous domain (NC-1). The seventh exon (exon 3, 178 bp) coded for sequences of the NC-1 domain. Five of the six -Gly-Xaa-Yaa- coding exons began with the second base coding for glycine, and only one exon began with a complete glycine codon at the 5' end. The results (i) suggest that the gene for the pro alpha 1(IV) chain of human basement membrane collagen is significantly larger than the genes for fibrillar collagens and (ii) show that it lacks the 54-bp exon repeats characteristic of fibrillar collagen genes.