Alternative splicing of fibroblast growth factor 1 (FGF-1) transcripts: a cellular dilemma in determining exon selection and exclusion.

During reverse transcription and polymerase chain reaction (RT-PCR) of kidney RNA using fibroblast growth factor 1 (FGF-1) gene-specific primers, we amplified five fragments with the sizes of 507, 410, 310, 285, and 216 bp. Cloning of the 507-, 310-, and 216-bp fragments revealed that the latter two share the same sequences to the 507-bp DNA but missing 197 and 291 bp, respectively. Characterization of the corresponding genomic DNA sequences showed that the three cDNA are alternative splicing products of the same gene. We further showed that these three transcripts are also present in brain in similar proportions albeit in a much lesser extent. Repeated attempts to clone the 410-bp fragment resulted in isolation of three additional cDNA clones; each has the inclusion of a distinct novel exon. Because each novel exon is delimited by the splicing donor and acceptor sequences, these cDNA clones are not likely to be RT-PCR artifacts. Yet their mRNA levels are extremely low, because we could not detect the corresponding PCR products on ethidium bromide-stained gels nor by Southern hybridization. These novel exons may be routinely used in different tissues. The nature of the 410- and 285-bp products was shown to be heteroduplexing among the 507-, 310-, and 216-bp PCR products. Identification of the three novel FGF-1 exons, which we designate exons -1B, -1E, and -1F, will facilitate the studies of the mechanisms of regulated alternative splicing.