The isolation of genomic 5′ regulating regions of genes starting from a suitable sequence (either a cDNA or an amino acid derived oligonucleotide) can be achieved in several ways. One of these, although long and labor-intensive, is the screening of a genomic library, followed by the isolation, subcloning, and sequencing of putative positive clones. A more direct approach is offered by inverse polymerase chain reaction (IPCR) using selfcircularized genomic DNA (1 ). Yet, PCR performed on circular DNA sometimes fails to produce an amount of amplification product sufficient to be visualized by EtBr-staining (2 ), thus hampering further isolation and cloning of the amplification product. The problem is usually overcome by relinearizing the circular DNA with a suitable restriction enzyme that has to cut in the region between the diverging amplification primers; such restriction site must not be present in the unknown flanking region, otherwise amplification will not take place.