A simplified genomic subtractive procedure for isolating deleted sequences.

New approaches have recently been developed for directly identifying genes without knowledge about their function and/or position in a genome (1–3). Among these, representational difference analysis (RDA) (2) is widely used in the detection of deleted DNA in cultured cancer cells (4). However, the procedure is complicated and time-consuming. First, preparation of the tester and the driver involves a series of steps including PCR amplification, digestion and recovery of PCR products from low melting agarose. It is well known that PCR products may not completely represent the original DNA sequences, especially in complex template mixtures. Furthermore, the low yield of recovered DNA from agarose can lead to further loss of representation. Secondly, single-stranded endonucleases may digest duplex nucleic acids due to non-specific activity when present in excess. It is, therefore, difficult to optimize the enzymatic conditions in RDA experiments. In order to overcome these difficulties, we have developed a modified RDA procedure (Fig. 1). The tester was made simply and rapidly by digestion of normal DNA with restriction enzymes; the driver from mutated DNA was sonicated to a size range of 500–1500 bp. Moreover, the single-stranded endonuclease (SI or Mung bean nuclease) was omitted while the enrichment efficiency was not affected (2,3). To test the feasibility and efficiency of our method, we adopted the same experimental model as Lysitsyn suggested. Human placental DNA with equimolar λDNA was used as the tester and the same placental DNA as the driver. In this experiment, λDNA was the target present in the tester but absent from the driver. The tester (0.5 μg) was cut with HindIII endonuclease and ligated to HindIII adapter I (primer I: 5′-AGCACTCTCCAGCCTCTCACCGCA-3′, primer II: 5′-AGCTTGCGGTGA-3′, 16 μM) (2). The reaction was carried out in a volume of 30 μl by overnight incubation with 2 U T4 DNA ligase (BRL) at 16 C. The driver DNA (40 μg) was sonicated to a size range of 500–1500 bp and mixed with the tester, then ethanol-precipitated, dissolved in 4 μl Tricine (Serva) buffer (pH 8.3), overlaid with 30 μl mineral oil and denatured at 100 C for 10 min. One microlitre of 5 M NaCl solution was added and then the mixture was reannealed for 20 h at 68 C. The resulting DNA was extracted with phenol–chloroform (1:1), precipitated with ethanol and redissolved in 30 μl sterile water. An aliquot (3 μl) of this solution was incubated with 1 U Taq DNA polymerase (Promega) in 50 μl of PCR mixture without primer for 5 min at 72 C to fill in the overhangs in the Figure 1. Schematic presentation of our procedure.