Characterization of the cathepsin B gene and multiple mRNAs in human tissues: evidence for alternative splicing of cathepsin B pre-mRNA.

We have cloned and characterized multiple messages for cathepsin B that differ in their 5' and 3' untranslated regions (UTRs) from human kidney and the hepatoma cell line HepG2. A comparison of these messages with the cloned human cathepsin B gene reveals that they arise by alternative splicing of a single gene. Processing at a cryptic intron donor site in exon 11 and splicing to exon 12 produces a 4.0-kb message with an alternate 3' UTR in addition to the 2.3-kb message described previously by Chan et al. (1986). Variable removal of exon 2 produces cathepsin B mRNAs which differ by 88 nucleotides in their 5'-UTRs. The ratio of the 2.3-kb to 4.0-kb transcript is about 2:1 in most of the tissues examined, but the ratio of mRNAs with variant 5' UTRs differs widely. Cathepsin B mRNAs lacking exon 2 are predominant in human tumors. In addition, human breast and colon carcinomas and a human melanoma contain a cathepsin B transcript that is also missing exon 3 encoding the signal peptide and 7 residues of the activation propeptide. An in vitro transcription/translation assay was used to demonstrate that this message could be translated from an internal methionine codon (residue 52), producing a 32-kD product lacking the signal peptide and more than half the propeptide. The transcription/translation assay also demonstrated that the variant messages differ in their rates of translation. The relative rates are about 8:2:1 for mRNA lacking exons 2 and 3 compared to mRNA lacking exon 2 and mRNA containing the full-length 5' end, respectively. These results suggest that the expression of cathepsin B in human tissues may be regulated in part at the level of mRNA processing.