2 GENES, COL4A3 AND COL4A4 CODING FOR THE HUMAN ALPHA-3(IV) AND ALPHA-4(IV) COLLAGEN CHAINS ARE ARRANGED HEAD-TO-HEAD ON CHROMOSOME 2Q36

We first isolated and characterized genomic DNA fragments that cover t he 5' flanking sequences of COL4A3 and COL4A4 encoding the human basem ent membrane alpha 3(IV) and alpha 4(IV) collagen chains, respectively . Nucleotide sequence analysis indicated that the two genes are arrang ed head-to-head. To determine transcription start site for COL4A4 gene , we performed RACE and RNase protection assays, indicating that there are two alternative transcripts presumably derived from two different promoters. Interestingly, one transcription start site (from exon 1') of COL4A4 is only 5 bp away from the reported transcription start sit e of COL4A3, whereas the other transcript (from exon 1) starts 373 nuc leotides downstream from the first one, generating the two kinds of tr anscripts that differ in the 5' UTR regions. Expression of these two t ranscripts appears tissue-specific; exon 1 transcript was expressed pr edominantly in epithelial cells, while exon 1' transcript showed rathe r ubiquitous and low expression. The nucleotide sequence of the promot er region is composed of dense CpG dinucleotides, GC boxes, CTC boxes and a CCAAT box but no TATA box. These results provide information to delineate the promoter activity for the tissue-specific expression of the six type IV collagen genes and basement membrane assembly in diffe rent tissues and organs. (C) 1998 Federation of European Biochemical S ocieties.