The human transthyretin (TTR) gene encodes a protein composed of four identical subunits with an important role in the plasma transport of thyroid hormone T4 and retinol. TTR spans 7.6 kilobases and consists of four exons. Two independent open reading frames (ORFs) with putative regulatory sequences have been described in the first and third introns, but their function--if any--is unknown. We have screened human cDNA libraries to determine if these sequences are transcribed. Transcripts of both ORFs were found in liver, pancreas and brain. Hybridization of the two sequences with multiple-tissue Northern blots further confirmed these results and revealed transcript sizes of approximately 1.5 and approximately 2.2 kb for ORF 1, and approximately 5.2 and approximately 7.8 kb for ORF 2. Rapid Amplification of cDNA Ends (RACE) was performed to characterize the full-length cDNAs containing each sequence. All products containing the ORFs were continuous in the genomic sequence corresponding to unspliced or partially spliced TTR. No evidence was found for novel transcripts containing productively spliced products of either ORF, or for shorter transcripts using the promoter and polyadenylation signals associated with them. ORF 1 RACE products identified in liver, pancreas and brain correspond to TTR transcripts in which intron 1 had not been removed; the transcripts containing ORF 2 may represent TTR hnRNA. Neither ORF is productively expressed as part of a larger transcript, or as an independent polypeptide.