L.M. Hinton

We have developed a strategy for the rapid enrichment and identification of cDNAs encoded by large genomic regions. The basis of this "direct selection" scheme is the hybridization of an entire library of cDNAs to an immobilized genomic clone. Nonspecific hybrids are eliminated and selected cDNAs are eluted. These molecules are then amplified and are either cloned or subjected to further selection/amplification cycles. This scheme was tested using a 550-kilobase yeast artificial chromosome clone that contains the EPO gene. Using this clone and a fetal kidney cDNA library, we have achieved a 1000-fold enrichment of EPO cDNAs in one cycle of enrichment. More significantly, we have further investigated one of the "anonymous" cDNAs that was selectively enriched. We confirmed that this cDNA was encoded by the yeast artificial chromosome. Its frequency in the starting library was 1 in 1 x 10(5) cDNAs and after selection comprised 2% of the selected library. DNA sequence analysis of this cDNA and of the yeast artificial chromosome clone revealed that this gene encodes the beta 2 subunit of the human guanine nucleotide-binding regulatory proteins. Restriction mapping and hybridization data position this gene (GNB2) to within 30-70 kilobases of the EPO gene. The selective isolation and mapping of GNB2 confirms the feasibility of this direct selection strategy and suggests that it will be useful for the rapid isolation of cDNAs, including disease-related genes, across extensive portions of the human genome.

We have developed modifications to direct cDNA selection that allow the rapid and reproducible isolation of low abundance cDNAs encoded by large genomic clones. Biotinylated, cloned genomic DNAs are hybridized in solution with amplifiable cDNAs. The genomic clones and attached cDNAs are captured on streptavidin coated magnetic beads, the cDNAs are eluted and amplified. We have applied this protocol to a 425kb YAC that contains the human IL4 and IL5 genes. After two cycles of enrichment twenty-four cDNAs were evaluated, all of which were homologous to the YAC. DNA sequencing revealed that nine cDNAs were 100% homologous to the interferon regulatory factor 1 (IRF1) gene. Six clones were 70% homologous to the murine P600 gene, which is coexpressed with IL4 and IL5 in mouse Th2 cells. The nine remaining clones were unique within the sequence databases and were non redundant. All of the selected cDNAs were initially present at very low abundance and were enriched by as much as 100,000-fold in two cycles of enrichment. This modified selection technique should be readily applicable to the isolation of many candidate disease loci as well as the derivation of detailed transcription maps across large genomic regions.