Improved blunt-end cloning by replacing EcoRV with Eco32I.

The restriction endonuclease EcoRV, which recognizes and cleaves DNA containing 5′-GATATC-3′ sequences, is among the best characterized and most frequently used bluntend cutting restriction enzymes (2). Its recognition site is present in many cloning vectors currently in use. Compared with that of cohesive ends, bluntend ligation is generally more difficult because a higher number of background colonies is formed that result from self-ligation of vector molecules (1). We (and others) have found that when using EcoRV even “white color” selection often fails so that one ends up with a large number of apparently positive (white) colonies. Here we demonstrate that this problem can be reduced if one uses Eco32I, an isoschizomer of EcoRV (3), instead. To compare the cloning efficiencies of EcoRV and Eco32I, we chose to insert a blunt-ended PCR fragment derived from the human β-catenin gene CTNNB1 into the GATATC site of vector pZErO 2.1 (Invitrogen, Carlsbad, CA, USA). The GATATC site of pZErO 2.1 is an integral part of an open reading frame encoding a lacZα/ccdB fusion protein. Insertion of a heterologous DNA fragment disrupts this open reading frame and results in both loss of β-galactosidase as well as rescue from the ccdB suicide gene activity. Thus, using pZErO 2.1, self-ligated vector molecules are efficiently selected against, thereby obviating the need for dephosphorylation of vector DNA before blunt-end ligation. Genomic DNA was isolated from human embryonic kidney-derived 293 cells, and a 675-bp DNA fragment derived from the human β-catenin gene CTNNB1 was amplified by PCR using primers bcat-fwd1 (5′-ACATTTCCAATCTACTAATG-3′) and bcat-rev1 (5′AAGTTCTGCATCATCTTGATAG-3′). The resulting PCR fragment contains multiple stop codons in all six possible reading frames, ensuring that inactivation of the ccdB suicide gene occurs upon insertion of the fragment in either orientation and any reading frame. The PCR product was gel-purified and extracted using Concert Gel Extraction System (Invitrogen) as recommended by the manufacturer. For cloning, the purified PCR fragment was made blunt-ended by treatment with T4 DNA polymerase and column-purified using High Pure PCR Product Isolation kit (both from Roche Applied Science, Basel, Switzerland). To generate vector suitable for ligation to the CTNNB1 insert, 500 ng pZErO 2.1 DNA were digested for 60 min at 37°C with 10 U EcoRV (Roche Applied Science) or 10 U Eco32I (MBI Fermentas, Vilnius, Lithuania) using supplied buffers. The digested vector was purified using the High Pure PCR Product Isolation kit. A 15-μL ligation mixture was then set up in a reaction containing 10 ng linearized vector, 10 ng PCR fragment (molar end ratio 1:5), 4 U T4 DNA ligase (Roche Applied Science) in 1× ligase buffer. A control reaction lacking PCR fragment was done in parallel. All ligations were incubated at 16°C for 16 h. One microliter of each ligation mixture was transformed into 50 μL chemically competent One Shot® Top 10 cells (Invitrogen), and the cells were spread on LB agar plates containing 50 μg/mL kanamycin. To determine the percentage of true recombinant clones, we prepared colony lifts using Biodyne® A filters (Pall, East Hills, NY, USA), which were subsequently hybridized to a radioactively labeled probe generated by amplification of an internal 224-bp fragment of CTNNB1 using primers bcat-fwd2 (5′-GATTTGATGGAGTTGGACATGG-3′) and bcat-rev2 (5′GCTACTTGTTCTTGAGTGAAGG3′). Hybridization-positive colonies were counted. We did not employ “white color” selection to check for the presence of inserts because in many previous experiments we found that self-ligated vector molecules lacking inserts may give rise to white colonies. No insert-containing colonies were obtained in the ligation reactions “vector alone” (Table 1). Ligation of the blunt-ended fragment into vector DNA cut by either enzyme, Eco32I or EcoRV, resulted in the same absolute number of insert-containing clones; however, when Eco32I-digested vector was used, a higher proportion of insertpostive clones (38%) was obtained as compared to EcoRV (15%). The difference found was highly significant (twosided P value < 0.001; Fisher’s exact test). In three independent experiments, the use of vector DNA prepared by restriction with EcoRV resulted in the formation of a significantly higher number of background colonies (Table 1). To investigate the possible reason for the observed difference in the formation of kanamycin-resistant colonies, we checked the integrity of the GATATC cloning site of clones obtained after cutting and simple self-ligation by sequence analysis. Ten clones each from EcoRV and Eco32I “vector alone” plates were picked, grown, and plasmid DNA was extracted using the High Pure Plasmid DNA Isolation kit. Sequencing reactions were set up following a dye terminator protocol (BigDye; Perkin Elmer, Boston, MA, USA), following the manufacturer’s instructions. Sequencing primers were standard M13 forward and reverse primers. Sequence analysis revealed a frequent single-nucleotide deletion within the vector’s cloning site that changed the GATATC motif to GAATC (GATTC in the complementary strand; Benchmarks