Answers are grounded in this paper and cite their sources.
The need to investigate a complex DNA sample consisting of unknown sequences still prompts the researcher to use in vivo cloning (cloning in bacteria, in phages and in other in vivo systems) as the only way to make the sample suitable for screening and sequencing. Here we demonstrate that it can be easily done by PCR, which saves a lot of effort. The main task of cloning—the amplification of individual molecules—is relatively easy to solve by PCR. It requires the ligation of adapter to all DNA ends in the sample, then dilution of the sample to produce a concentration at which the volume being taken for amplification contains about one molecule, and PCR (‘cloning takes’) with a primer which corresponds to the adapter. However, in this case the amplification products would have the same flanking sequence at both ends and therefore could not be sequenced using these flanks as primer annealing sites—two texts would be read simultaneously. We have developed a technique which, though being close to the aforementioned simple scheme, ensures that the final amplification product is flanked by different adapter sequences at the different ends. Such fragments can be sequenced easily from either of the ends. In our in vitro cloning procedure, we ligate a pair of pseudo double-stranded adapters ( 1 ) in a mix to all DNA molecules in the sample. Each adapter is a pair of oligonucleotides of uneven length complementary to each other at the 3′-end of the longer oligo ( Fig. 1 A), forming a structure that is able to participate in ligation as a normal blunt end. However, only the longer oligo is ligated, because the 5′-ends of both oligos, including the shorter one, lack phosphate groups. The DNA stretch complementary to the ligated longer oligo is produced before the first PCR cycle at temperatures when the smaller oligo already falls off its binding site, but longer DNA molecules still retain their double-stranded structure (in fact, it occurs while the sample is heated prior to the first denaturation stage). The adapters should be rather long (about 40 bp) for two reasons: (i) to make it possible to perform two rounds of PCR with nested primers, which is known to be much more powerful and specific than a one-round amplification; (ii) to suppress the amplification of molecules which had obtained identical adapters at both ends during ligation. Only molecules asymmetrically flanked by the two adapters can be amplified, because after the denaturation of DNA strands the complementary ends of each strand of symmetrically flanked molecules anneal to each other, hiding the primer annealing site before the primer can bind to it. This effect works very effectively even for molecules as long as 5 kb, completely inhibiting the amplification of symmetrically flanked molecules, as was reported recently ( 1 , 2 ). We have already made use of it in several novel PCR-based techniques ( 1–5 ). The reliability of the ligation-dilution-amplification scheme as proposed for the amplification of individual molecules has been verified in a model system. The DNA test sample was Hae III treated φX174 DNA. Figure 1 B shows that after the ligation of adapters such a dilution ratio can be selected which produces on average one band per cloning take picked out at random from the original φX174/ Hae III band pattern. We have proved that such a product is asymmetrically flanked by amplifications with oppositely directed primers specific for one of the φX174/ Hae III bands and adapter-specific primers (data submitted but not shown). The asymmetry of the resulting PCR products proves once more that they were amplified from a single molecule: otherwise, the bands almost certainly should have contained fragments of opposing asymmetry (as adapters' combinations ‘AB’ and ‘BA’ for the same molecule type are equally probable during ligation) and no asymmetry should have been detected in the described experiment. It should be stressed that true ‘single molecule per reaction’ aliquots in the dilution scheme appear statistically only in ∼30% of all takes, even when the dilution ratio is ideally determined. Other aliquots would contain either more than one PCR product or be empty. Therefore, we consider it reasonable to deal with dilutions producing about 10 molecules per cloning take. In this case, statistical fluctuations hinder the analysis much less. After amplifcation a small number of products in each take can be easily discriminated by electrophoresis. The bands on the gel can be treated as individual clones quite similar to bacterial ones: they can be screened and excised from the gel for separate re-amplification and direct sequencing from adapter-specific primers. Any method designed for sequencing of uncloned PCR products is suitable (for example, 6,7), because the in vitro clones are pure PCR products. In our routine practice we have never encountered any problems with sequencing them. As an example of such an approach, Figure 2 shows pictures of differential screening of a subtracted cDNA sample (prepared as described in 4). The sample was cloned in vitro as described above and a panel of some 50 clones were screened for sequences not present (or present in a much lower concentration) in the reference sample. The band that gave differential signal was picked out and reamplified. Its special character for the subtracted sample was confirmed by dot hybridization of this band as a probe to both samples—the subtracted and the reference ones (data not shown). Finally, the band was successfully sequenced using a fmol sequencing kit (Promega) ( 8 ). ( A ) Structures of suppressive adapters and amplification primers used in this work. ( B ) Amplification of single molecules from the φX174/ Hae III sample. After Hae lII treatment DNA was precipitated with ethanol and NaAc, dissolved in water and 70 ng sample were taken for adapter ligation. Suppression adapters were taken in final concentration of 2 µM in a total ligation mix volume of 10 µl. Ligation was allowed to proceed for 2 h at room temperature. Then the ligation mixture was diluted sequentially in the solution containing 10 mM Tris-HCl pH 8.0, 1 mM MgCl 2 (to prevent diluted DNA denaturation in low-salt conditions) and 50 µg/ml tRNA (to stabilize highly diluted samples). One microliter of each dilution corresponding to DNA concentration of 0.001–0.1 fg/µl (which for φX174 corresponds approximately to 1–100 molecules) was taken for a two-round PCR (described below) to determine exactly 10 molecules/µl dilution (data not shown). As the determination was finalized, a total mix for the first round of PCR was prepared containing 2 µl of this dilution in a PCR mixture of total volume 20 µl. It was subdivided into 20 aliquots of 1 µl each and applied to two-round amplification as described below. The products of this amplification were analyzed by agarose gel electrophoresis. Lanes 1–20, different cloning takes of one molecule/reaction volume dilution; lane 21, original φX174/ Hae III (with adapters ligated) band pattern. All amplifications were performed by KlenTaq-Pfu (30:1). Amplification reactions contained 1x PC-2 buffer (50 mM Tris-HCl pH 9.1, 16 mM ammonium sulfate, 2.5 mM MgCl 2 , 150 g/µl BSA), 250 µM dNTPs, 0.3 µM of each primer 25 U KlenTaq/100 µ1 reaction mixture and 2 µl TaqStart antibodies (Clontech) per 100 µl reaction mixture for a hot start. PCR was carried out in a Hybaid OmniGene with temperature control in the reaction tube. The first round was performed with primers A1 and B1 as follows: 95°C for 2 s, 68°C for 15 s, 72°C for 2 min, 25 cycles. Then the resulting PCR mixtures were diluted at a 1:50 ratio with water and 1 µl of each was applied for the second round of PCR in a total volume 20 µl. It was performed with primers A2 and B2 as follows: 95°C for 5 s, 67°C for 15 s, 72°C for 2 min, 25 cycles. Our in vitro cloning approach is much easier, cheaper and less time consuming than in vivo cloning: it takes 2 h for the ligation of adapters, one PCR to determine the required dilution ratio and a full-scale PCR to obtain a panel of clones like that shown in Figure 2 . There are no intervening procedures like bacterial transformation and DNA isolation and purification; a researcher always deals with DNA of sufficient purity for sequencing. Hybridization signals from such DNA during screening are much more reproducible as compared to those typical for bacterial colonies or phage plagues. The latter property makes in vitro cloning especially useful in differential screening procedures, when the responses of each clone to several probes are compared. An example of differential screening of subtracted library cloned in vitro. In vitro cloning was performed as described in the legend to Figure l, but a ligation mixture dilution was selected that gave about 5 molecules per cloning take. ( A ) Agarose gel electrophoresis of in vitro clones. Lane 1, original cDNA sample; lane 2, 100 bp ladder; lanes 3–13, different cloning takes. ( B ) Results of hybridization of the panel shown in (A) (lanes 3–13) with the subtracted library as a probe. ( C ) Hybridization of the same panel with a reference library as a probe. The band hybridizing specifically with the subtracted library is marked by an arrowhead. We thank our colleagues from Clontech Laboratories, Inc. Paul Siebert and Alex Chenchik for encouraging this work, Luda Diachenko for adapters' design. We are also grateful to Prof. Eugene Sverdlov for his special attention and discussions. This work was supported by the Russian Foundation for Fundamental Research (grant no. 94-04-11257) and International Science Foundation (grant no. MOJ300).
No takes yet. Share an insight, caveat, or question.
Konstantin A. Lukyanov (1996) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: