outcross of integrated strains

I got one integrated strain using gamma irradiation, which has 100% transmission rate, all the progeny have GFP.
To eliminate the mutations caused by irradiation, I cross this strain with N2 male.
But after cross, the progeny all don’t have GFP, no matter male or hermaphrodite. Does anyone know what’s wrong with the outcross?

I’m unclear: are you saying that after outcrossing none of the progeny have GFP (“all don’t”), or are you saying that after outcrossing you’re unable to establish a line that only throws GFP animals (“don’t all”)?

Because if it’s the former (GFP disappears) I don’t have an explanation to offer (well, I have a silly idea, but it’s not something I’ve heard of ever happening and I don’t know how it would). On the other hand, if it’s the latter (can’t recover apparently homozygous lines after outcrossing) I can offer a couple of suggestions:

  1. A homozygous mutation is synthetic lethal with loss of the transgene, which remains extrachromosomal. This depends a lot on the transgene, and the co-injection marker, but it can definitely happen if you used lin-15 as a co-injection marker. Note that after outcrossing the transgene will behave more like an extrachromosomal array again (transmission won’t be strictly Mendelian, you might see more mosaicism)
  2. Balanced lethals. If your transgene is integrated, but linked to a lethal mutation, and the other copy of that chromosome has a tightly linked lethal mutation of its own, you could get the behavior you observe. This is a bit more plausible than it might at first seem (because, the odds seem steep), as gamma is liable to rearrange chromosomes and suppress recombination, so the two linked lethals don’t have to be incredibly close to each other physically. You should be able to spot this in the outcross, as only half of cross-progeny should have the transgene. Also, you should see dead eggs on the non-outcrossed strain.
  3. Best-case scenario: your transgene is integrated, and is homozygous-viable, but your picking strategy is biased against recovering homozygotes - for example, integrant homozygotes are slow growing or have some other subtle phenotype that isn’t immediately obvious when the strain is growing on its own but dramatically lowers the odds that animals you pick will be homozygous. This can be somewhat more likely if the integrant is in a reciprocal translocation (all else being equal, transgene-positive animals would have a 1/5 chance of being homozygous instead of 1/3).

Unfortunately, it’s the former one.
Although there are some progeny have GFP, but those worms are from self-fertilized.
All the males and most of the hermaphrodite are without GFP.
Also the original strain is very sick, growing slowly.

And your original integrant animals form a vulva?

I’ve got a model for this, sort of, unfinished. But it’s ugly, and I’m not sure it quite works:

  1. There’s a complicated genomic rearrangement, with the transgene integrated onto something that’s either subject to meiotic loss (a free duplication) or not homozygous.
  2. This fragment containing the transgene, whatever it is, dominantly causes severe sickness and slow growth.
  3. The fragment containing the transgene complements a lethal mutation elsewhere in the genome, or otherwise restores something resembling euploidy or suppresses a lethal phenotype, possibly by balanced lethals.
  4. When you mate in, you only really see the animals that don’t inherit the transgene - you no longer need it to prevent lethality/aneuploidy/whatever, and animals with it are really sick and slow growing.

Mind you: this would mean the transgene-containing animals would still be present in the outcross, but you just wouldn’t spot them. Or maybe, because of the dominant slow-growth phenotype, you wouldn’t recognize them as cross-progeny - especially if the genomic rearrangement involving the transgene involved the X chromosome somehow.

It’s not a great model, and I haven’t worked out the kinks. And it has a big drawback: it really doesn’t offer any insight or recommendations beyond the obvious, ie chucking this candidate integrant strain and starting over.

The crosses seem to generate cross-progeny, unless your theory is that the crosses are contaminated with wild-type hermaphrodites, which doesn’t seem likely to happen more than once.

i like hillel’s balanced lethal ideas. also, your integrated strain might have undergone physiological/epigenetic compensation for dominant deleterious effects of the transgene (or it could be homozygous for a recessive suppressor).
the F1s that carry the transgene are then going to be particularly unhappy.
sounds fairly nightmarish in any case.
i would scrap the strain and do it by MosSCI or miniMos.

I agree that the strain you are currently working with sounds like more trouble than it’s worth. I would redo the integration.

Alternative methods (MosSCI, Crispr/Cas9, miniMos) would also be good approaches to try if the repeat gives unsatisfactory results.