Back in the 1930s, scientists noticed that water fleas and rats fed restricted diets while young went on to live longer. The same effect has since turned up in fruit flies and mice. What we did not know was how a meal eaten in infancy could affect health weeks, months, or even years later. A recent Nature study might have found the answer.
A team led by Fumiaki Obata, a biologist at the RIKEN Center for Biosystems Dynamics Research in Kobe, Japan, found a protein that carries a record into adulthood of what fruit flies ate as larvae, influencing how long they live.
Hungry maggots
Lab fruit flies are fed a mix of yeast and sugar, with yeast being their main source of protein. “We decreased only the yeast concentration in the diet, from eight percent to either one or two percent,” Obata explains. This low-protein diet was introduced roughly halfway through the larval period. Once the flies emerged from their pupae as adults, they went back on standard food.
In Obata’s experiment, the protein-restricted flies, males and females alike, outlived their well-fed siblings. But it came at a cost. The flies were paler and lighter, weighing sometimes 28 percent less. The females also laid fewer eggs.
“It’s very common, actually. Reproduction and lifespan are always in a tradeoff relationship,” Obata says. “This is also the case in this early-life dietary restriction. They have fewer eggs and they are slightly smaller. But they have a lifespan extension.”
This typical tradeoff wasn’t the whole story, though. When the team added amino acids, the building blocks of proteins, back into the low-yeast larval food, the lifespan boost disappeared. Somehow, the adult flies’ bodies were remembering how much protein they had eaten as larvae. Obata and his colleagues called this a nutritional memory and set out to find where exactly in the body that memory was kept.
Nutritional memory
Comparing gene activity in adult flies raised on low protein and standard diets turned up around 100 candidate proteins that could store the information about early protein intake. “We basically went one by one, checking which would be important,” Obata says.
To narrow the search down, the team tagged the larval food to tell which proteins in an adult fly’s body were built from food it ate as a larva. Larvae were raised on a synthetic diet in which two amino acids, lysine and arginine, were made with rarer, heavier isotopes of carbon and nitrogen atoms. Once the larvae became adult flies, the team switched to food with a second set of amino acids tagged with different, lighter isotopes.
Then the researchers used mass spectrometry to weigh fragments of proteins from the flies’ heads. The same protein fragment might register as a few units heavier or lighter, depending on whether it was built from larval or adult food. “This is basically only achievable with this stable isotope experiment,” Obata says.
Three days into adulthood, nearly 64 percent of the proteins in the flies’ heads were still made from amino acids eaten by larvae. By day six, that only dropped to about 46 percent. Prominent among them were ribosomal proteins, which are part of the complex that makes all other proteins. Flies fed a protein-restricted diet as larvae had fewer of these and made new proteins more slowly during their first week of adulthood.
This immediately raised a question. Flies go through metamorphosis inside a pupal case, and pupae don’t eat. So the larval amino acids that ended up in adult ribosomal proteins had to be stored somewhere during metamorphosis, and whatever held them had to be sensitive to how much protein the larvae ate.
The team went looking for a protein that both responded to the larval diet and survived into adulthood. And they found it.
Storage protein
It was a protein called larval serum protein 2, or Lsp2. Larvae stockpile it as a pantry for metamorphosis when they stop eating. “The flies express this Lsp2 massively in early life,” Obata says. “And we knew that there is some carryover of this Lsp2 protein into the adult stage, and that Lsp2 is responsive to dietary protein in the early stage. That led us to think that this would be the best candidate in the end.”
Protein-restricted larvae made less Lsp2, and the levels stayed low well into adulthood, even after the flies went back to eating normally. Researchers found that low Lsp2 in larvae apparently tells the adult body to keep producing less of it. When the team genetically silenced Lsp2 in larvae, the adult flies ended up with fewer ribosomal proteins, slower protein production, and longer lives—just like the flies raised on the low-protein diet.
The team also noticed that Lsp2 is unusually rich in two amino acids, phenylalanine and tyrosine. Removing tyrosine from the larval food, or cutting phenylalanine to a quarter of its usual level, was enough to lower Lsp2 and extend lifespan; restricting isoleucine, an amino acid that doesn’t show up frequently in Lsp2, did nothing. “The low-tyrosine, low-phenylalanine flies cannot make enough [Lsp2],” Obata explains. “It’s the same thing that happens under the low-protein diet condition.”
The full chain of events Obata’s team reconstructed starts with less larval protein, which leads to less Lsp2, which means fewer ribosomes and slower protein production in young adults—which results in a longer life. But not all links in this chain are equally solid.
Missing links
What’s most firmly established today is the final step. “If you have less translation and less protein production, that leads to the lifespan extension,” Obata says. This lifespan extension, he argues, relies on improved proteostasis, the cell’s ability to keep its proteins in good working order.
The link between Lsp2 and ribosomes is less understood. “What we have to find out is why this Lsp2 is preferably going to the ribosome,” Obata says. “This part is still kind of a mystery.” The team also couldn’t directly demonstrate that fewer ribosomes cause the longer lifespan by restoring ribosome levels. “Ribosomal proteins are a mixture of roughly 100 different proteins,” Obata told Ars. “If you want to manipulate ribosomal levels, you have to either up or down these 100 proteins at once, which is not technically possible at this moment.”
But even with all the details figured out, feeding longevity diets to toddlers is unlikely to become a thing in the foreseeable future, mainly because mammals, including humans, have no direct equivalent of Lsp2.
Translation issues
Obata thinks the best functional counterparts of Lsp2 in humans are albumin and globulins, which together make up around 90 percent of the protein in our blood. Like Lsp2, albumin is made in large amounts, turns over quickly, and is sensitive to how much protein we eat.
Testing whether these proteins carry a nutritional memory just like Lsp2 does would be a stepwise process. “We have to do this first in mice, because there we can manipulate proteins directly, and then maybe in primates,” Obata says. “In humans, we would have to perform studies looking at whether there is a correlation between the level of albumin and lifespan.” If albumin gets confirmed as a human nutritional memory carrier, the team will still have plenty of other details to figure out.
Human albumin is not particularly rich in tyrosine or phenylalanine, so if a similar mechanism exists in people, it may rely on different amino acids. “We have to find out which amino acids are key, which molecular carrier would be equivalent to Lsp2, and then whether there really is a connection between lifespan and the diet in people,” Obata says.
Nature, 2026. DOI: 10.1038/s41586-026-11031-3
