Cloning PCR products using TA vectors

Cloning PCR products using TA vectors

by Paul N. Hengen, Ph.D.

Methods and reagents is a unique monthly column that highlights current discussions in the newsgroup bionet.molbio.methds-reagnts, available on the Internet. This month's column compares some of the commercially available vectors for cloning PCR fragments and discusses some of the problems encountered with them. For details on how to partake in the newsgroup, see the accompanying box.

To clone the blunt-ended, double-stranded products of the polymerase chain reaction (PCR) into plasmid vectors, it is necessary to perform blunt-end ligation, which are less efficient than cohesive-end ligations. To circumvent this problem, the molecule can be modified by some other means to create a sticky end on it. One common method is to engineer restriction enzyme sites within each of the primers, which are then incorporated into the PCR product. After amplification, the DNA is digested with the appropriate restriction enzymes and ligated into the multiple cloning region of a linearized plasmid vector.

A little T and A
Another way is to create complimentary tails of single-stranded DNA on the ends of both the plasmid and the insert with terminal deoxynucleotide transferase (TdT) and either deoxyadenosine triphosphate (dATP) or deoxythymidine triphosphate (dTTP).

A peculiar property of Taq DNA polymerase is that, when DNA fragments are generated by PCR, the enzyme introduces one or two extra deoxynucleotides onto the 3'-end of blunt double-stranded DNA in a template-independant manner. Any one of the four nucleotides can be added when a mixture of all of them is present in the reaction, but for unknown reasons preference is given to incorporation of dATP [1]. T-tailing of vector DNA can therefore also be accomplished by generating T-overhangs on any blunt-ended vector using Taq polymerse and dTTP alone.

Interestingly, this `terminal transferase-like' phenomenon may be partly responsible for the failure of PCR when primers are stored for extended periods at -20 degrees C, as was described in a previous Methods and reagents column (see TiBS 20, 42-44). If the 3'-end of one of the primers confers some loop back, or primer-dimer structure, repeated freezing and thawing cycles may allow for secondary structures to be formed within or between primers over time.

During PCR, aberrant modification of the 3'-end could occur as soon as the enzyme touches the solution while held at low temperature and this could cause the specificity of the primer to change. The altered primers may then cause primer-dimers or misprime elsewhere, and squelch out the fragments of interest. Modification followed by mispriming could be the sole factor responsible for the squelching out of reaction components leading to PCR failure.

On the other hand, several commercially available kits for cloning PCR fragments take advantage of the ability of Taq polymerase to cause 3'-modifications. Most of these use a plasmid vector with thymidine residues linked onto the 3'-ends of linearized plasmid DNA, which would allow some annealing to occur between the vector and the A-tailed PCR product to be ligated. These vectors are generally referred to as T-vectors, and the process called TA cloning.

Vector Selection
Rae Nishi of Oregon Health Sciences University ( nishir@ohsu.edu ) recently wrote about her experience using T-vectors to clone a PCR fragment, and a discussion arose regarding the pros and cons of several vectors which can be used for this, including the pBluescript derivative pCR-Script [TM] SK(+) cloning kit from Stratagene, pCR [TM] II supplied with the TA Cloning [TM] kit from Invitrogen, pT7Blue T-Vector kit from Novagen, pGEM[R]-T from Promega, and the SureClone [TM] ligation kit from Pharmacia. Some netters have reported that some of these vectors have been giving them problems, and a few ways to avoid common pitfalls were suggested.

For example, the pCR-Script [TM] vector used for blunt-end cloning has been found to give high white-to-blue-ratio background problems in the absence of added insert, while the pCR [TM] II has a replication region of pBR322 and a relatively low copy number compared to the others of pUC origin, which can be a disadvantage. Some suggested that the newer version of this plasmid is more useful than the original because it contains a kanamycin esistance gene as well as an ampicillin resistance gene. Although it should not be much more work to make up different media for the experiment, the new vector allows previously poured plates to be used, and it is easier to compare with other plasmids on the market containing the same genetic marker system. However, despite costing more than ampicillin, kanamycin reduces the chance of mistakenly picking a satellite colony that could have no insert DNA. In addition, the INValphaF'-competent Escherichia coli cells provided within the Invitrogen TA cloning [TM] kit were said to grow poorly in selective LB media. Many people who bought this kit have used a different host from that supplied.

Instability of the T-ends is an important consideration, and the Novagen pT7Blue T-Vector may be less likely to lose Ts from the 3'-end, as has been found with vectors of this type, when repeatedly frozen at -20 degrees C and thawed. Although it is not clear how this has been achieved, Novagen claim that the chemistry used to create and/or protect the T-ends provides a more stable plasmid vector. It may simply be that there is less exonuclease activity in the plasmid preparation.

Dr Nishi settled on the Promega plasmid pGEM[R]-T over the others because of the vector's excellent reputation for reliability. This plasmid is suitable for cloning PCR fragments because it is actually pGEM[R] -5Zf(+) DNA digested with EcoRV and single 3'-T overhangs added.

With plasmid and PCR fragment in hand, however, she encountered several unforeseen problems that could have occurred with any one of the T-vectors. For example, the first batch of plasmid seemed to work very well, but all the other batches had problems. One lot had an unacceptably high background of white colonies when the plasmid alone was used to transform E. coli. Another lot seemed to be missing the T-overhang so that even the control insert would not ligate.

Later, a more serious problem was encountered - two clones with different inserts were each isolated by a different person in the lab and found to be missing the T7 end of the vector. That is, none of the enzymes which should recognize sites within the multiple cloning site (MCS) on the T7 end of the insert would digest the DNA. After checking that the enzymes were working properly, it was concluded that a spontaneous deletion had occurred within the MCS.

Another netter had also found that some of the presumed recombinant plasmids had major deletions somewhere near their MCSs. This concerned only white colonies growing on media containing X-gal and never the faintly blue ones, both types of which were identified as PCR fragment inserted within the vector. Over several experiments, it was observed that the proportion of deletion-containing plasmids seemed to vary between lots.

It appeared that either the PCR fragments or the T-vectors do not have the correct overhanging nucleotides, and that some batches of vector break down more easily than others, perhaps having been contaminated with exonuclease. Furthermore, if only one end is ligated while the other remains blunt, a deletion is more likely to occur on only one side of the MCS, as was observed.

It was suggested that to avoid selecting deletions, both white and light blue colonies should be screened for the insert by PCR, using primers at least 100 base pairs away from the MCS.

Some of the problems with the artificial multiple cloning sites have been seen previously by others. For instance, while constructing several subclones within a pUC derived plasmid, I have lost restriction sites that were originally within the MCS, and sometimes I could detect small segments deleted from the vectors used. The vectors with large MCSs may therefore be somewhat unstable, possibly because they contain several small palindromic sequences linked closely together.

On the other hand, other netters have had few problems and, upon sequencing the recombinant plasmids, found that the only modification to the MCS was the T-tailed insert. In any case, when selecting a vector for TA cloning, it may be a good idea to avoid unnecessarily large MCSs and select a vector based on a few cloning sites specific for the designed experiments.

Another idea is that batches of plasmid DNA may contain more than one type of plasmid vector which, upon transformation, yield a high proportion of white colonies. If the DNA is sequenced from the T7 side, no insert DNA is found and the proper restriction sites are missing from the MCS.

One way around these problems would be to transform the host strain immediately after having obtained the vector and to select out a blue colony for restriction analysis. After confirming the structure of the vector, plasmid DNA should be isolated for the cloning experiment and digested with EcoRV, the enzyme used for creating the linear vector with T-overhangs. Unfortunately, it is becoming more evident that buying a kit is not an adequate substitute for good clean science.

Anti-kit netters have several slick techniques up their sleeves for creating their own T-vectors. Instead of buying expensive kits, many create a batch of T-tailed vector for their own purposes. For example, the plasmid pDK101 [ATCC 77406], available for little or no cost is a derivative of plasmid pGEM[R]5fZ(+) which, when cut with the endonuclease XcmI, provides the T-overhangs needed for TA cloning [2,3]. Another way is to digest pBluescript with EcoRV and perform a Taq polymerase 3'-terminal addition in the presence of dTTP. [4]

References

[1] Clark, J. M. (1988) Nucleic Acids Res. 16, 9677-9686

[2] Kovalic, D., Kwak, J.-H., and Weisblum, B. (1991) Nucleic Acids Res. 19, 4560

[3] Mead, D. A. et al. (1991) Bio/Technology 9, 657-663

[4] Marchuk, D., et al. (1991) Nucleic Acids Res. 19, 1154