Libraries of cDNA clones are valuable resources for analysing the expression, structure, and regulation of genes, as well as for studying protein functions and interactions. Full-length cDNA clones provide information about intron and exon structures, splice junctions and 5'- and 3'-untranslated regions (UTRs). Open reading frames (ORFs) derived from cDNA clones can be used to generate constructs allowing expression of native proteins and N- or C-terminally tagged proteins. Thus, obtaining full-length cDNA clones and sequences for most or all genes in an organism is critical for understanding genome functions. Expressed sequence tag (EST) sequencing samples cDNA libraries at random, which is most useful at the beginning of large-scale screening projects. However, as projects progress towards completion, the probability of identifying unique cDNAs via EST sequencing diminishes, resulting in poor recovery of rare transcripts. We describe an adapted, high-throughput protocol intended for recovery of specific, full-length clones from plasmid cDNA libraries in five days.

The production of comprehensive cDNA clone collections is an important goal of the human and model organism genome projects. cDNA sequences are used to determine the structures of transcripts, including splice junctions, polyadenylation sites, and 5' and 3' untranslated regions (UTRs). cDNA collections are also valuable resources for functional studies of genes and proteins. Expressed Sequence Tag (EST) sequencing is the method of choice for recovering cDNAs representing a majority of the transcripts encoded in a eukaryotic genome. However, EST sequencing samples a library at random, so it realizes diminishing returns as the project progresses. To drive cDNA collections toward completion new methods are needed to recover cDNAs representing specific genes and alternative transcripts, including transcripts with low expression levels. We describe a simple and effective inverse-PCR-based method for screening plasmid libraries to recover intact cDNAs for specific transcripts. We tested the method by screening libraries used in our Drosophila EST projects for 153 transcription factor genes that were not yet represented by full-length cDNAs. We recovered target-specific clones for 104 of the genes: 46 exactly match, 30 improve and 28 partially match current gene annotations. Successful application of the screening method depends on cDNA library complexity and quality of the gene models. The approach should be effective for improving cDNA collections for other model organisms and the human. It also provides a simple and rapid method for isolating cDNAs of interest in any system for which plasmid cDNA libraries and complete or partial gene sequences are available.