1Present address: F. Hoffmann-LaRoche AG, CH-4002, Basel, Switzerland. Although the polymerase chain reaction (PCR) (1) provides a powerful method for amplifying DNA from extremely small amounts of starting material, blunt-end cloning of DNA fragments generated by PCR has proved to be difficult, even after repairing the ends with T 4 DNA polymerase or Klenow fragment. A reasonable level of success has been achieved by incorporating restriction endonuclease sites at the ends of the primers; however, this approach is only useful when the sequence of the fragment is known. The observation that DNA polymerases can add a single nontemplated residue (with a distinct preference for adenosine) to dsDNA (2) have led to a number of general cloning methods (3-6) taking advantage of the single 3' A overhang. All of these methods rely on vectors having single T overhangs generated either by enzymatic tailing of the linearized vector (3's) or by cleavage with the restriction enzymes HphI (6) or XcmI. (4) Analysis of all 11 restriction enzymes generating a single-base overhang showed that in all cases the efficiency of ligation and subsequent recutting was less than 95%. This is compared to the 32 enzymes with either two or three base overhangs where, in all except three cases, the efficiency of ligation and subsequent recutting was greater than 95% (New England Biolabs Catalogue, 1990-1991). It may well be that there is a connection between the ability of DNA polymerases to add a single nontemplated base and the poor performance of ligase to join these overhangs. Ligase could recognize a single-base overhang as an error as it knows that single bases are nontemplated additions by DNA polymerases. There is suggestive evidence for this in that DNA cut with the restriction enzymes BstNI (CC'WGG), PleI (GAGTCNNNN'N), ScrFI (CC'NGG), and Tth111I (GACN'NNGTC), which all generate single-base overhangs cannot be religated following cleavage (New England Biolabs Catalogue, 1990-1991). Bearing this in mind, we decided to design a cloning procedure that would take into account the A overhangs generated in the PCR and also minimize the problems associated with ligating single base overhangs. We have done this by electing to use a tailing procedure to obtain PCR fragments having extended single-stranded overhangs of two to three adenosine residues. Thus, we have designed a compact cassette containing two copies of the recognition sequences for XcmI and PflMI and a single recognition sequence for BcgI. Cleavage with these enzymes generates either one, two, or three T overhangs following the removal of an internal fragment (Fig. 1). In addition, we have flanked these restriction sites (which generate the T overhangs) with MscI sites to enable the complete cassette to be removed by bluntend digestion. The internal fragment also contains a Sinai site that can be used to reduce background due to uncut vector if necessary. The cassette was constructed by synthesizing an oligonucleotide that contained the coding sequence for the cassette as well as flanking regions containing restriction sites PstI and BamHI. The oligo was made doublestranded, cleaved with PstI and BamHI, and ligated into pBluescript II SK+ (Stratagene, La Jolla, CA). The insert is 48 nucleotides long, maintaining the reading frame of the 13-galactosidase gene allowing blue/white selection of recombinants. A naturally occurring BcgI site in the ampicillin gene of the pBluescript vector was removed using site-directed mutagenesis with the oligonucleotide (5'-GTGTATGCGCCTACCGAGTTG-3') while maintaining the amino acid sequence of the ampicillin resistance gene. Vectors were prepared by digestion of the plasmid DNA with the appropriate restriction enzyme and elution from low-melting-point agarose. A 764-bp fragment prepared by PCR was eluted from a 1.5% low-meltingpoint agarose gel. The fragment was polyadenylated in a 50-pA reaction [100 mM K cacodylate pH 6.6, 12.5 mM Tris/ HC1, pH 7.0, 125 i~g/ml BSA, 0.1 mM DTT, 1 mM CoC12, 20 I~M dATP, 10 units of terminal deoxytransferase (Boehringer Mannheim), 0.5-1 pmole of 3'-OH termini] for between 3 and 16 min at 37~ The reaction was terminated by the addition of 0.5 M EDTA pH 8.0 (2 I~l) and then extracted twice with an equal volume of phenol/chloroform and once with an equal volume of chloroform prior to ethanol precipitation. Polyadenylated fragments were ligated into prepared vector in 10-1~1 reactions (50 mM Tris/HC1, pH 7.8, 10 mM MgCl2, 1 mu DTT, 1 mM ATP, 100 i~g/ml BSA, 20 ng of vector, and 0.8 Weiss units of T 4 DNA ligase) and incubated overnight at 16~ The ligation mix was transformed into E. coli XL-1 Blue competent cells and plated