These results show that a gene or DNA fragment of interest can be subcloned from one entry vector to an expression vector as easily and efficiently as when using recombination-based systems. However, in contrast with those systems, the recombinant plasmids obtained with our strategy do not contain a recombination site at the junction sites. In fact, the only requirement is 4 nucleotides of defined sequence at the restriction/ligation site. This means that only one amino-acid is fixed at the ligation site (if this region is part of translated sequences). Moreover, the sequence of the 4 nucleotides can theoretically be freely chosen by the experimenter since it can consist of any 4 nucleotide sequence of choice. However, sequences to avoid for these 4 nucleotides are the 16 palindromic sequences, which would reduce cloning efficiency, therefore leaving a choice of 240 possible sequences (out of 256). At present all sequences that we have tested have performed well, but more use of this cloning strategy will reveal whether this can be generalized for all 240 possible sequences and for any specific combination of these.
This cloning protocol also allows cloning of at least three DNA fragments from three separate entry clones into an acceptor vector. The only requirement is to design a set of compatible restriction sites in the entry clones and the acceptor vector.
As with recombination-based systems, the cloning method described here can be used to transfer one gene of interest (or more generally any DNA fragment of interest) from one entry clone to a series of different expression vectors designed to have compatible cloning sites. For example, we have made several compatible vectors that allow expressing the same gene of interest in either viral vectors or in standard non-replicating expression vectors ().
Model for subcloning of a gene of interest into a library of compatible expressions vectors.
One advantage of recombination-based systems is that restriction enzymes can be avoided both for generation of the entry clones, for example using the ligation-independent cloning strategy of the In-Fusion system from Clontech, and for subcloning into expression vectors using site-specific recombination. In contrast, one potential limitation of the protocol presented here might come from the occasional presence of one or several internal BsaI site(s) in the gene of interest. Since the BsaI restriction site has a 6 base pair recognition sequence, BsaI restriction sites are expected to be present in average every four kb for genomes with a GC content of 50%. In practice, the restriction site frequency will of course depend on the base composition of each specific organism. For example, out of 86 annotated coding sequences in the Arabidopsis thaliana chloroplast genome (GenBank NC_000932), only six contain one BsaI site and four contain two sites. Nevertheless, there are at least three possible solutions for bypassing this problem. One solution consists of performing the digestion and ligation sequentially, and heat inactivating the restriction enzyme before performing the ligation. A variation of this solution is to perform the restriction-ligation for 5 minutes, then heat inactivating the enzymes, then adding fresh ligase to the mix, and ligating for another 5 to 10 minutes. We often use this strategy when the insert to be cloned contains an internal BsaI restriction site. This is extremely efficient since the restriction-ligation stably ligates both insert fragments at the ends of the vector in the first step, and the following ligation step only needs to religate a linear molecule. A second solution would consist of using a second type IIs enzyme (for example BpiI) to perform the cloning from entry clone to expression vector. This would however require having a second set of expression cloning vectors with restrictions sites for the second type IIs enzyme, and therefore, this is not a preferred solution. The third and more general solution consists of eliminating any BsaI restriction site from the fragment of interest upon cloning in the entry clone, as was shown in the results section. Elimination of the BsaI site(s) from an entry vector requires sequencing the amplified sequence to confirm that no PCR-derived mutation is present; however, sequencing is also necessary for all other entry clones since they are all made by PCR. The extra work consists only of designing an additional pair of primers for each BsaI site to be removed.
For the method to work, it is also necessary to remove any BsaI restriction site in the expression vector, except for the two sites flanking the LacZ alpha fragment (or any other negative selection marker that an experimenter would prefer to use).
In summary, the cloning strategy described here, that we call ‘Golden Gate’ cloning, combines the convenience and efficiency of currently used recombination-based cloning systems with unique insert precision.