Goro Yoshizaki and his colleagues at the Tokyo University of Marine Science and Technology had previously shown that male salmon could be injected with cells from closely-related trout to produce viable trout sperm.
When the sperm were introduced to trout eggs, healthy trout offspring were produced.
Now, the researchers have taken the work a step further, and showed that salmon could not only be the biological fathers but also the mothers of trout offspring.
The researchers injected trout spermatagonia - the early, stem-cell stage of the sperm - into salmon embryos, so that the growing salmon produced trout sperm and eggs.
If trout spermatagonia are injected into normal male salmon embryos, the fish will grow up to produce a mix of different types of sperm - some salmon, and some trout.
In the team's previous work, when sperm from male salmon treated with primordial germ cells (an even earlier stage of sperm) were used to fertilize trout eggs, just 0.4 percent of the resulting offspring were healthy trout. The rest were hybrids that did not survive.
To increase this percentage, the team turned to salmon designed to have three sets of chromosomes instead of the usual two, making them sterile.
When they were injected with spermatagonia, the only viable sperm the males later produced came from these injected cells - making them 100 percent trout.
The scientists met with success doing the same with sterile female salmon. When the female fish were injected as embryos with spermatagonia, the eggs they produced were all trout eggs.
When mated, these salmon parents produced healthy trout offspring, which in turn mated to give a healthy second generation, said Yoshizaki.
David Penman, a fish geneticist at the University of Stirling, UK, said, the technique could be very useful for storing back-up genetic material of different fish species that are today under threat, because spermatagonia can be easily cryopreserved.
'The fact that an early form of sperm could be used to prompt the growth of eggs shows just how flexible these fish are to this type of treatment, It's really amazing to see that they injected spermatagonia and got eggs — the fish are very plastic,' said Penman.
'The problem with gene-banking when you come to fish, is that it's almost impossible with eggs. The eggs are very big, very yolky, which makes them nearly impossible to freeze.
'If this technique is broadly applicable to other fish, it will mean that their eggs don't need to be preserved — they could be made at a later date, in a surrogate fish,' he said.
The study appears in the journal Science, reports Nature.