Several PCR-based methods are available for walking from a known region to an unknown region in cloned or uncloned genomic DNA. The methods are of three types: inverse PCR (1), randomly primed PCR (2) and adaptor ligation PCR (3-6). However, these methods have not been generally applied to walking in uncloned genomic DNA because they are either complicated or inefficient. Recent improvements to these methods have been applied to uncloned genomic DNA (7, 8), however, walks have been limited to distances of< 1 kb. We have investigated the application of'long and accurate PCR'(9, 10) to walking in uncloned genomic DNA. We initially examined the use of'unpredictably primed PCR'(3), a new method based on randomly primed PCR and'vectorette PCR'(5), which is based upon adaptor ligation. As a model system we attemptedto walkupstream fromexon 1 of the human tissue-type plasminogen activator (TPA) gene which has been previously characterized (11).

In our hands both'unpredictably primed PCR'and'vectorette PCR'generated multiple PCR products even after nested PCR was perforned. Although upon analysis some of the PCR products were found to be derived from successful walks, the presence of multiple PCR products would complicate further characterization without Southern blot hybridization or extensive cloning. We have improved upon the adaptor ligation method by combining'vectorette PCR'with a newly developed method termed'suppression PCR'(12). A special adaptor is ligated to the ends of DNA fragments generated by digestion of human genomic DNA with EcoRV, Scal, Dral, PvuII and SspI separately. Following adaptor ligation, a small amount of the DNA is used as a template for PCR using adaptor primers and gene specific primers. The enzymes used were selected because they have six-base recognition sites and generate blunt ends. The sequences of the walking adaptor and primers are shown in Figure 1. One end of the adaptor is blunt so that it will ligate to both ends of any DNA fragment generated by restriction enzymes that yield blunt ends. The adaptor also contains two rare restriction enzyme sites for NotI (staggered ends) and SflI/SmaI (blunt ends) to allow cloning into commonly used vectors such as pBluescript (Stratagene).