In the past, analyzing the structure and function of a genome has been a dicey process. Once a mutation is observed and a gene is identified as being responsible, biologists will knock out that gene and introduce different versions of it to see what phenotypes they can induce. But existing ways of reintroducing the missing DNA back into a genome has suffered from two major problems: First, the transgene often ends up in random parts of the genome, where the adjacent DNA can modulate its expression--a phenomenon known as a position effect. "That's problematic," explains geneticist Hugo Bellen, the study's lead author, "you are starting to compare apples and oranges." He notes that sometimes a phenotype may emerge after introduction and other times no effect may be seen, but that scientists can't be sure why this is happening. In addition, researchers have trouble inserting and manipulating large pieces of DNA (bigger than 30 kilobases), because of the bacteria used as a vector for this DNA. Specifically, the bacteria incorporate the genes into circular DNA molecules called plasmids, which carry the genetic segments into the fly's genome. The bacteria, however, make many copies of the plasmids; with so many copies around they often start to recombine "and you don't know what you've got anymore," Bellen notes.
According to Koen J. T. Venken, a graduate student whom Bellen credits as the impetus and leader of the study, the new method developed by the Baylor team primarily came about to combat the size limitations of the previous technology. Venken first sought a vector that could handle large amounts of DNA efficiently, settling on a bacterial artificial chromosome, which was designed a couple years ago and is known to maintain only a few copies of the DNA (though it can produce a high number of copies if it is induced). Placed into a bacterium, the chromosome integrates into the bacterial vector--a process known as "recombineering." "This methodology is very important because it allows you to very quickly integrate almost any piece of DNA in these vectors," remarks Bellen, who says this technology also allows scientists to put single point mutations in a gene and then reinsert the gene into a genome as well as tagging a segment of DNA with fluorescent markers.
Once the vector is ready, the Baylor team wanted to be able to guide it to a certain part of the genome every time, so they turned to a technique called ΦC31, which was known to work in human and mouse cells. Bellen and Venken attached artificial docking sites to phages--a virus that targets bacteria--which they introduced in both the bacteria and in the fruit fly genomes to act as sticky points for the DNA segment being implanted in the latter's genome. "It seems like the whole Drosophila community is switching over to ΦC31 because it's both more efficient and its 100 percent site specific," notes Michele Calos, a geneticist at Stanford University who is a pioneer in work on the ΦC31 integration scheme in mammalian cells. This study, she adds, "will encourage people to use ΦC31 for larger fragments in other organisms."
Thomas Kaufman, a biologist at Indiana University, praises Bellen's team for overcoming both the size limitations and the localization specificity of previous vector-based DNA introduction. "It opens up another compartment of the genome to analysis," he says. "It's a step up in the transgenic technology that's available in this model system that opens up new pathways of analysis." Bellen adds that among other things, the method eliminates the need to reintroduce the same transgene over and over again and look for an average effect: "Now you need only one transgene, because if you inject into the same site, you'll always have the same expression level and the same position effect. So, you're not comparing apples and oranges anymore, you're comparing the same gene with subtle changes." Bellen claims the new methodology will work for DNA segments up to 150 kilobases in length--the fruit fly genome is made up of an estimated 14,000 genes, only 10 of which, according to Bellen, are too large to be manipulated with P[acman]. He also believes the technique will work reasonably well for studying the structure and function of proteins in most other model organisms, including laboratory mice.