Biologists have long known that new protein-coding genes can arise through the duplication and modification of existing ones. But some protein genes can also arise from stretches of the genome that once encoded aimless strands of RNA instead. How new protein genes surface this way has been a mystery, however.
Now, a study identifies mutations that transform seemingly useless DNA sequences into potential genes by endowing their encoded RNA with the skill to escape the cell nucleus—a critical step toward becoming translated into a protein. The study’s authors highlight 74 human protein genes that appear to have arisen in this de novo way—more than half of which emerged after the human lineage branched off from chimpanzees. Some of these newcomer genes may have played a role in the evolution of our relatively large and complex brains. When added to mice, one made the rodent brains grow bigger and more humanlike, the authors report this week in Nature Ecology & Evolution.
“This work is a big advance,” says Anne-Ruxandra Carvunis, an evolutionary biologist at the University of Pittsburgh, who was not involved with the research. It “suggests that de novo gene birth may have played a role in human brain evolution.”
Although some genes encode RNAs that have structural or regulatory purposes themselves, those that encode proteins instead create an intermediary RNA. Made in the nucleus like other RNAs, these messenger RNAs (mRNAs) exit into the cytoplasm and travel to organelles called ribosomes to tell them how to build the gene’s proteins.
A decade ago, Chuan-Yun Li, an evolutionary biologist at Peking University, and colleagues discovered that some human protein genes bore a striking resemblance to DNA sequences in rhesus monkeys that got transcribed into long noncoding RNAs (lncRNAs), which didn’t make proteins or have any other apparent purpose. Li couldn’t figure out what it had taken for those stretches of monkey DNA to become true protein-coding genes in humans.
A clue emerged when Li’s postdoc, Ni A. An, discovered that many lncRNAs have a hard time exiting the nucleus. The researchers used a sophisticated computer program to identify differences between protein-coding genes whose mRNA got out of the nucleus and the DNA sequences that produced RNAs that did not. The program homed in on stretches of DNA known as U1 elements, which when transcribed into RNA make the strand too sticky to make a clean escape. In protein-coding genes, these elements have mutations that make the RNA less sticky. So, for an lncRNA to escape the nucleus and give its instructions to a ribosome, the parental DNA must acquire those key U1 mutations or somehow make that transcribed section get cut out of the RNA strands altogether.
“This makes perfect sense because for an RNA to be translated, it needs to go the cytoplasm [where ribosomes are found] first,” says Maria Del Mar Albà, an evolutionary biologist at Hospital del Mar Medical Research Institute.
Li’s team scoured the human and chimpanzee genomes for de novo protein-coding genes that had lncRNA counterparts in rhesus monkeys, as well as the crucial U1 element mutations needed to exit the nucleus. Eventually they came up with 45 exclusively human genes and 29 genes shared by humans and chimps that fit the bill. Next, the researchers homed in on nine of these protein genes that are active in the human brain to see whether they could learn what each was doing. Li’s collaborator Baoyang Hu, a neuroscientist from the Chinese Academy of Sciences Institute of Zoology, grew clumps of human brain tissue called cortical organoids with and without each of these genes and identified two that made the organoids grow slightly bigger than normal.
When Hu introduced one of these genes into mice, their brains also grew larger than normal and developed a bigger cortex, the wrinkly outer layer of the mammalian brain that in humans is responsible for high-level functions such as reasoning and language. The second gene did likewise in mice, and also caused the animals’ brains to develop more humanlike ridges and grooves. Those mice performed better on tests of cognitive function and memory than mice lacking this gene, the team says it will report soon in Advanced Science.
Overall, the findings suggest these de novo human genes “may have a role in brain development and may have been a driver of cognition during the evolution of humans,” says Erich Bornberg-Bauer, an evolutionary biophysicist at the University of Münster.
Manyuan Long, an evolutionary biologist at the University of Chicago, calls the new study “a breakthrough in the understanding of the molecular evolutionary processes that generate [new] genes.” In an indication of how widespread those processes may be, Long’s group has found that most of the recognizable de novo genes in rice were once lncRNAs, and that lncRNAs also helped form new genes in bamboo. But she is more cautious about interpreting the role of de novo genes in brain evolution. Organoids are far simpler tissues than the brain itself, she notes, and human and mouse brains have evolved along very different paths.
Xiaohua Shen, a molecular biologist at the Tsinghua University School of Medicine, adds that she wishes the authors had studied a larger sample of mice to be sure the differences in brain size from the gene additions couldn’t be explained by natural variation.
The work suggests profoundly influential de novo genes might arise through subtle changes in their DNA sequence, Carvunis says, but there’s still much to be learned about how escaped lncRNAs eventually become true genes. “There are a lot of barriers to gene birth,” she says. “I hope this work will contribute to inspiring more research towards understanding what these barriers are and how emerging genes can overcome them.”