Analysis of the human expressed sequence tag (EST) database identified four clones that contain sequences of previously uncharacterized genes, members of the ATP-binding cassette (ABC) superfamily. Two new ABC genes (EST20237, 31252) are located at Chromosome (Chr) 1q42 and 1q25 respectively in humans, as determined by FISH; at locations distinct from previously mapped genes of this superfamily. Two additional clones, EST 600 and EST 1596, were found to represent different ATP-binding domains of the same gene, ABC2. This gene was localized to 9q34 in humans by FISH and to the proximal region of Chr 2 in mice by linkage analysis. All genes display extensive diversity in sequence and expression pattern. We present several approaches to characterizing EST clones and demonstrate that the analysis of EST clones from different tissues is a powerful approach to identify new members of important gene families. Some drawbacks of using EST databases, including chimerism of cDNA clones, are discussed.

We have isolated a human cDNA encoding a novel ATP-binding cassette (ABC) protein whose gene was previously localized to chromosome 1q42 [Allikmets et al. (1995) Mamm. Genome 6, 111-117]. The gene transcript is expressed in all human tissues examined, with the highest levels in bone marrow. A non-expressed pseudogene also exists at chromosome 15q13-14. The new protein, which is most similar to the mitochondrial (M)-ABC1 protein, was also localized to mitochondria and therefore designated 'M-ABC2'. The N-terminus of M-ABC2 was shown to contain a mitochondrial-targeting signal sequence.