Site-directed mutagenesis by fusion of contiguous DNA fragments.

Efficient site-directed mutagenesis (SDM) is a powerful tool for the analysis of the function of DNA sequences. The polymerase chain reaction (PCR) (4,9) and commercially available methods for in vitro selection of mutated unmethylated DNA (Stratagene, La Jolla, CA, USA) have been used increasingly as rapid alternatives to classical methods of site-specific mutation (11). Generating mutations at the termini of PCR-amplified fragments by incorporating mismatches at the 5′ ends of priming oligonucleotides is usually an uncomplicated procedure. However, mutation of specific sites that are internal to the flanking primers is more difficult. The most widely applied PCR procedure for SDM has been the megaprimer method (2). The approach involves two sequential PCR amplifications. Double-stranded (ds)DNA produced from a flanking and mutagenic primer during the initial reaction is used as a megaprimer for the second amplification step with the opposite flanking primer. A number of modifications of this method have been described (5,6,8,10) to overcome poor yields of the final PCR product resulting from variable megaprimer efficiency. The double-stranded nature of the megaprimer diminishes its effectiveness as a primer for the second PCR. Also, the addition of residues to the 3′ ends of PCR fragments that are not template dependent (1) might result in mismatches and inhibition of DNA polymerase extension during the second amplification reaction. An uncomplicated method for SDM that is based on standard methods of PCR, restriction digestion, ligation and plasmid cloning is described here. Two contiguous DNA fragments are amplified with two pairs of primers. The internal primers introduce the intended mutation(s) at the sequence where the amplified DNA fragments adjoin each other (Figure 1). Hexanucleotide recognition sequences of blunt-cutting restriction endonucleases are incorporated to permit ligation of the fragments at the exact point of their contiguity. Restriction digestion also enables removal of extra 3′ residues that might have been added during PCR. After fusion, the six nucleotides at the point of ligation in the mutated sequence comprise two trinucleotides derived from each of the two blunt-cutting restriction sites. Thus, the sequence requirement to carry out the procedure is the presence of two adjoining trinucleotides in the mutated DNA, each of which includes the 5′ half and the 3′ half of the recognition sites of two different blunt-cutting restriction enzymes. There are presently 20 restriction enzymes available commercially (New England Biolabs, Beverly, MA, USA) that digest symmetrically at individual palindromic hexanucleotide sequences to produce blunt ends. The probability of a random trinucleotide comprising the appropriate half of a blunt-cutting recognition sequence is thus approximately 0.3 (20 restriction sites out of 64 possible trinucleotide sequences). The probability of two adjoining trinucleotides constituting a point of fusion is approximately 0.1 (0.3 × 0.3). On average, in a randomly chosen sequence, such a hexanucleotide should be found every 15 bp (15 bp comprise 10 overlapping hexanucleotides). Moreover, the site of a single mutagenic primer mismatch is on average a maximum of 7 bp from the closest possible ligation point. This is sufficiently close to the 5′ end of a standard length primer (for example, a 30-mer) to allow incorporation of a noncomplementary 5′ end to generate a restriction site, while at the same time permitting matched hybridization at the 3′ end to allow DNA polymerase extension. Statistically therefore, it is unlikely that the number of nucleotides between the site of the intended mutation and the point of fusion is too great for a PCR primer to incorporate a restriction site at the 5′ end and matched hybridization at the priming 3′ end. We applied the approach to mutate three specific bases in a template sequence encoding the RNA encapsidation signal (ε) of the hepatitis B virus (HBV) AYW subtype. The mutated HBV DNA generates pre-genomic HBV RNA of the ADW subtype and with a mutation commonly found in ε, which is associated with a HBV e antigen-negative phenotype among chronic HBV carriers in sub-Saharan Africa (8). In our hands, attempts to generate the mutations using the megaprimer method were unsuccessful.