I have been trying to integrate a simple unc-122p::GFP array after microinjection by using the protocol on WormBook.
I have repeated this twice, each time singling around 150 fluorescent progeny of the UV-irradiated worms. Starting from
a strain with around 20% transmission, I have a couple of plates with much higher transmission, say 80 or even 90%. I
thought these may be heterozygous integrants, so I singled 10 fluorescent progeny from 10 of these plates. However,
this never leads to a homozygous strain, so presumably the integration wasn’t successful, but how is it possible that they
suddenly show such an increased transmission nonetheless? And what could I do differently to get a proper integration?
Thank you for your advice!
Are you using UV alone or UV with the sensitizer, trimethyl psoralen. I recommend the latter, protocol here:
https://theolb.readthedocs.io/en/latest/misc/tmp-uv-integration.html
We would typically try to pick ~300 transgene-positive F1’s from the integration. From every one of these F1 plates that throw trans gene-positive progeny, we would then pick 2-4 transgene-positive F2 animals. You can keep one plate per original F1. We would typically get 2-6 integrants per 300 F1’s picked.
That doesn’t look like a great protocol. There is nothing relevant or valuable about incomplete “high transmission”; if the transgene has been integrated into the genome, there’s four simple possible outcomes:
- The integrated transgene is homozygous, and viable. 100% transmission.
- The integrated transgene is heterozygous, and neutral (no disadvantage for homozygotes). You’d expect 75% transmission; over multiple generations, would converge on looking like 50% transmission.
- The integrated transgene is heterozygous, and is recessive-lethal. You’d expect 67% transmission in one generation, compounded with each generation.
- The integrated transgene is heterozygous, as part of a reciprocal translocation. If homozygotes are viable, you’d expect 87% transmission in the first generation, 80% if they aren’t, and it gets annoying to calculate in subsequent generations (the progeny classes from the heterozygote are 1 homozygous wild type : 4 heterozygous : 1 homozygous translocation : 10 aneuploid).
As you can see, the only outcome that resembles 80-90% transmission is the reciprocal translocation, giving you a reagent that acts as a balancer for about a quarter of the genome. So: “increased transmission” is arguably an interesting result, but it doesn’t necessarily get you any closer to the integrated transgene you want. To be sure, 80-90% transmission could be your impression of (or just random chance acting on) a simple Mendelian 75% - but your goal is a healthy homozygote, and 100%, not spotting a heterozygote and 75% transmission, and getting lucky. Aiming for “high transmission” is not a good strategy. In any case, what drives transmission frequency for an extrachromosomal array is not well understood (and 20% is quite low); rather than select a transgene to integrate on the basis of its transmission frequency, I would recommend finding one whose expression you like, whatever its transmission frequency, and then doing what you must to get the integrant. Though, this can fail: if expression is highly dose-sensitive, then it could be a problem that a homozygous integrant will have twice the dosage of the extrachromosomal array.
There are two popular protocols to find integrants. One is the method kevinem recommends (pick hundreds of F1s after irradiation each to their own plate, from each plate put two F2s each on its own plate, look for plates with homozygous integrants). It is the older method, and is maybe the gold standard, the most certain to succeed, making the fewest assumptions. It’s also extremely labor-intensive and uses a lot of plates. My favorite version is the old “Koelle Lab Protocol”, which doesn’t seem to be directly online anymore but can still be seen navigating his old site at the Wayback Machine.
The other method is to let the transgenic line starve out a couple of times (without any sort of selection for transgenics), then look for homozygous integrants. I associate this with Oliver Hobert, but can’t find a protocol from him. This version in WormBook looks fine. This is certainly the most labor-efficient method, when it works.
Thank you both for your fast and thorough replies! Perhaps part of my problem is sheer numbers; many of these protocols seem to suggest singling 200-400 animals and I’ve only been doing 150. I believe the method I am using is based upon Oliver Hobert’s protocol, I leave the irradiated animals on their plates for two weeks to starve before singling fluorescent animals (so not exactly the one on WormBook, I was given a protocol by a colleague). Maybe I should do so in cycles of starving and feeding, as opposed to a long starvation period.
This protocol always worked well for me and requires less picking than most:
http://wormlab.caltech.edu/documents/10-extrachromosomal_array_integration_protocol.pdf
The advantage of letting them starve out is that it basically enriches for continued transmission and gives a few generations in which an integrated transgene can be homozygosed. This is why it’s usually done in bulk, as in the protocol I linked and in the protocol Barth linked. I suppose there’s no reason not to combine it with singling out hundreds of F1s, but it won’t be as labor-efficient that way.
In any case, the usual estimate is that following gamma irradiation you might get about 1 integrant for every 300 or so F1s put on individual plates. Remember, from each F1 in that protocol you’d look at 2 individual F2s, giving you about a 50/50 chance of recovering a homozygosed integrated transgene if it happened. My own experience is pretty close to that number. Unfortunately, I don’t know how well it compares when UV is used instead. Assuming UV is no better than gamma for this purpose, to not get an integrant from 150-200 F1s (or, indeed, from 3-400) would not be at all surprising. If you are going to single out F1s, I wouldn’t bother unless you were doing at least 600.
Just an update for anyone who is interested, I repeated the integration twice in parallel, using two different protocols, and each eventually yielded two successfully integrated strains. Both started with 0.030 J/cm² UV irraditation, followed by:
- Singled 540 F1’s, later singled 2 F2’s from each F1 (sometimes more if tranmission was high), scored F3 progeny for 100% transmitting strains.
- Starved irradiated plates, chunked onto new plates and starved again, singled 100 worms and scored their progeny for 100% transmitting strains.
Thanks for the help!