Place your bets please!

A big part of my new project is to generate a null mutant of two genes in Drosophila to see what effect they have. My hope is that it kills them or has some nice developmental phenotype I can then dissect further- we have proteomics, from human cells, suggesting this should be the case.

So the reason we make a mutant is fairly straight-forward. If you take a gene away and something goes wrong then you can claim that gene is required to prevent those things going wrong or more accurately you can infer what the gene is required for in a normal situation. There are basically 3 ways you can make a mutant in flies.


The old school way is to mutagenise them by any means necessary such as X-rays but the preferred method is by Ethyl Methanesulphonate (EMS) which introduces random mutations in the fly. The problem here is that it's random and mutates any and everything. There are tricks to make it more specific to your gene of interest but this technique is probably better suited to screening, where you find an interesting phenotype and then work out which gene was mutated.

The second method is imprecise P-element excision. P-elements are transposable elements of DNA which can move around the genome. In some cases the presence of a P-element in a gene is enough to mutate the gene as you've just stuck a large bit of DNA into it that isn't of any use to the resulting protein. Even if the P-element doesn't cause a mutation it can still be used to generate a mutant though. That's because when a P-element is mobilised and it leaves it's genomic location it can occasionally take a chunk of the neighbouring DNA with it. If you're lucky it takes away enough of your gene of interest to make it a null. The drawbacks with this technique is that it's random  there's no way to guarantee you'll get the mutation you want so you have to scale things up quite a lot and even then it may not work - or delete more than one gene.

The third technique relies on homologous recombination, which is basically how very similar stretches of DNA can swap genetic material between themselves. We can take advantage of this by creating a stretch of homology flanking the gene of interest but lacking the actual gene. It is then possible for a recombination event to take place where you swap the DNA from a wild type fly with the DNA that lacks your gene of interest, thereby creating a null mutation. Of the three techniques this method allows the finest level of control as to what you are deleting.


So you'd think I'd just go ahead and do the homologous recombination technique? Well, there a a couple of drawbacks with the method too. Mainly it can be an absolute pain to get the homology arms from the PCR cloned into the vector. That vector then needs to be injected and incorporated into flies before you even begin the homologous recombination and then you may not be guaranteed the recombination will occur. So the main issue is time and a little bit of troubleshooting with the cloning. It's possible to have performed a P-element excision a lot quicker and while it's random if it generates a null mutant it's just as valid. But P-excisions are random.

From previous experience I was unable to even get past the PCR stage with homologous recombination as I simply couldn't generate 3-4Kb homology arms in that region no matter how many oligos and super polymerase i tested. Then again I've done 4 P-element excisions three of which were done by slaves students. One of them failed another gave a mutant which to this day doesn't really make much sense as the P-element is still there but now has the correct phenotype. The last one worked.

My attitude this time was to do both things at the same time as to be honest I have the time for it and there's no point in saying "I should have done the other one" 3-4 months later. Plus as I'm doing this for two genes and a student is doing it for another one I'm thinking this is a good way to test which is the best method for me. I'm saying for "me" because the results may not be conclusive but we scientists can be fairly superstitious at times so the one that works (or disappoints the least) will probably be the favourite even if it fails the next 5 times.

In the "race" so far I'd say the homologous recombination method is winning in that I am at the cloning of the homology arms into the vector stage. The P-element method is still warming up though as i had to check the p-element was there (it is!) and I'm currently isogenising them - a fancy word for ensuring the genetic background is as close to identical on the chromosome of the gene of interest.
Things will probably change this month though as the P-excision gets underway. We shall see.

EDIT: I'll put some diagrams to illustrate the methods better later on. Just need to see if there are any figures I can borrow...