In a nutshell:
- Extremely efficient with frequencies approaching 70%
- At predetermined locations: one can obtain transgenes at a desired location
- Directly comparable results: Because the position of the transgene is controlled, all constructs integrated into the same locus can be directly compared.
- Economical: Because identical transgenic lines can be remade easily, there is no need to keep all the transgenic stocks, just the integration site and the integration vector.
With all of the genetic tools in Drosophila, it is not surprising to find that there are a number of strategies one can use to create transgenic flies. However, by far, the most common of these procedures is based on the use of P-elements.
In our lab, P-element mediated transformation is quite efficient. Generally speaking, once the P-element transformation vector is made and injected with a helper plasmid containing the P-element transposase, we obtain transformation frequencies in the range of 5-15% (meaning 5-15% of the surviving injected embryos give rise to at least one transgenic offspring). This frequency is quite reasonable, as long as experimenters have no preference as to where the transgene is located, or how well the transgene is expressed. This is because P-element mediated transformation is, for the most part, random. The P-element transgene can land almost anywhere in the genome, even in an essential gene! Also, depending on where the P-element lands, it sometimes comes under the influence of nearby enhancers or silencers, which can and do affect a transgene's expression (a phenomenon called position effect). Taking these factors into account, even when an experiment requires only a generally specific location and expression level, the actual efficiency of P-element mediated transformation is quite a bit lower. For example, to obtain a homozygous viable transformant where the transgene is located on the second chromosome, would cut your frequencies by more than half.
The randomness of P-element transformation makes direct comparisons between the activities of two or more constructs nearly impossible. The only way to make comparisons is through statistical analysis, often requiring many independent transformant lines for each construct. Not only does this require more tedious rounds of injections, but also requires a lot of extra work in the characterization and maintenance of these lines.
The key to the PhiC31 integrase system is non-random integration. Isolated from the bacteriophage PhiC31, the PhiC31 integrase (frequently also written as: ΦC31 integrase) encodes a serine-type recombinase that mediates the sequence-specific recombination between two largely different attachment sites, called attB and attP, which share a 3 bp central region, where the crossover occurs (Thorpe et al., 2000).
Because the reaction catalysed by the PhiC31 integrase is site-specific, one can pre- select a location in which to integrate, thereby eliminating the randomness of transgenesis. All that would be required for this are a selection of attB or attP sites in the genome (landing platforms) and a plasmid containing the complementary site (an integration vector). Because these genomic integration sites would be pre-characterized, if someone wanted a viable and highly expressed transgene on the left arm of chromosome 2, he or she would simply have to choose a line with those characteristics. What's more, if this person wanted to compare the activity of this construct to a mutated version of the same construct, he or she would simply have to integrate the mutant construct into the same landing platform. Because everything should be the same between these lines except the integrated construct, the results are directly comparable, and could easily be repeated or expanded upon by anyone with the same integration vector and landing platform.
How it works:
Because the two sites recognized by the PhiC31 integrase differ and the recombination process in essence mixes the two sites into two different sites (attR and attL), &C9B8;C31 based integration is unidirectional.
attB + attP + integrase → attR + attL + integrase
What this means on a practical level is that if you have an attB site on a plasmid and an attP site in the genome, the PhiC31 integrase will only mediate the integration of the plasmid into the genome and not its subsequent excision, since the resulting attR and attL sites can not serve as functional substrates for the integrase. This differs from the more commonly used FLP and Cre recombinases that each mediate recombination between two identical sites (FRT and loxP sites, respectively). Thus, because the sites are identical, recombination between the two sites recreates the sites again. This means the reaction is completely reversible. In fact, because after integration the two sites become located on the same DNA molecule, the excision reaction is highly favored. People have tried to circumvent this problem through the use of mutated FRT and loxP sites or by using two different recombination sites (such as a loxP and an FRT site) to mediate a cassette exchange. Some of these methods have worked quite well and we advise you see Horn and Handler (2005) and Oberstein et al. (2005) for more details. Recently, a recombinase-mediated cassette exchange (RMCE) technique based on the phiC31 integrase, has been presented by Bateman et al. (2006).