Jenste: As far as I know GFP in axolotls is not sex-linked. It would be annoying to work with, so the scientists wouldn't breed any that came up that way.
![Smile :) :)](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
So no, male or female shouldn't make a difference.
Michael: Yes, it can get more complicated in practice. Based on what you describe, there are two reasonable scenarios.
1. Multiple insertion loci of GFP in the axolotl genome. We obviously have stable transformants (since they can have GFP offspring), so that means the GFP gene is integrated into their DNA somewhere. If there were one place it was inserted, GFP would act like any ordinary gene controlling a single trait, and we'd see the perfect Mendelian segregation I was assuming in my previous response. However, if there are two different places in the genome that GFP is present, you could get something resembling a dihybrid cross. (
Dihybrid cross - Wikipedia, the free encyclopedia -- the green square in the figure would be the only axolotl without GFP.) And there might be even more than two. In my experiments I make multiple lines of any transgenic, so different GFP loci almost certainly exist, and I have no idea how or how many times GFP axolotls have reached hobbyists.
2. GFP sometimes gets silenced. I see this in my transgenics all too often. Even though the gene is inserted stably in the genome, sometimes the cells can "catch on" that there's something that doesn't belong there. We don't know exactly how it works in every organism, but you can see where it would be really useful for resisting viral infection: if a virus comes in and starts making a ton of copies of itself, the organism whose cells can respond by shutting down production of that virus will be a lot healthier than one that doesn't "notice". So you'd stop getting a protein made from the silenced gene, and you wouldn't see GFP any more. Usually the silencing is heritable, so the offspring would carry GFP but also carry the cellular "memo" not to make any.
Without doing crosses specifically to evaluate these hypotheses, it's hard to tell what's going on for sure, and there could easily be elements of both.